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 ) 91 — 9 6

An investigation of the chemical synthesis and

high-temperature sintering behaviour of calcium
hydroxyapatite (HA) and tricalcium phosphate
(TCP) bioceramics
A. C U® N E Y T T AS, * , F. K O RKU S UZ ‡ , M . T I M UC, I N * , N. A KKA S, $
*Department of Metallurgical and Materials Engineering, $Department of Engineering
Sciences, ‡Health Center, Middle East Technical University, Ankara 06531, Turkey

The experimental conditions for the synthesis of sub-micrometre, spherical particles of

calcium hydroxyapatite [Ca10(PO4)6(OH)2] (HA) and tricalcium phosphate [Ca3(PO4)2 ] (TCP)
are investigated through chemical coprecipitation from the aqueous solutions of calcium
nitrate and di-ammonium hydrogen phosphate salts. The precipitation process employed
was also found to be suitable for the production of sub-micrometre HA/TCP composite
powders in situ. The synthesized pure HA and TCP powders were found to be stable even at
1300 °C in air for prolonged heating times. Bioceramic sample characterization was achieved
by powder X-ray diffraction (XRD), scanning electron microscopy (SEM), energy dispersive
X-ray spectroscopy (EDX), and density and surface area measurements. Crystallographic
analyses of HA powders were performed by the Rietveld method on the powder XRD data.

1. Introduction obtained by precipitation from Ca(NO ) and

32
Ceramics are historically the oldest of the synthetic (NH ) HPO solutions and sintered above 800 °C,
42 4
materials and their medical use can be traced back to such as a- and b-TCP. Jarcho et al. [15] have pre-
the Egyptians. In Egyptian sarcophagi, both filled and viously shown in polycrystalline HA ceramics ob-
artificial teeth have been found in the mouths of mum- tained by precipitation from Ca(NO ) and
32
mies. According to a dentistry book from the late (NH ) HPO solutions, that at higher sintering tem-
42 4
1700s, dental implantation and transplantation were peratures such as 1250 °C (for 1 h) HA samples would
already popular in the eighteenth century [1]. In 1892, decompose according to:
Dreesman published a report on the use of plaster of
Paris to fill bone defects [2]. Almost 30 years later, Ca (PO ) (OH) P2(Ca (PO ) )#Ca P O #H O
10 46 2 3 42 4 2 9 2
Albee and Morrison were the first to report on the use (1)
of tricalcium phosphate (TCP) to fill bone gaps [3].
However, it was not until the 1960s that research into and documented this with an electron micrograph
the application of ceramics as bone replacements in- depicting the formation of a second phase at the HA
creased, with osteoceramic (a composite ceramic of grain boundaries in sintered samples.
TCP and MgAl O ) being one of the first to be pro- Jarcho et al. [16], in a later study, claimed that
2 4
posed, and used successfully in experiments on dogs the synthesis of ‘‘pure’’ b-tricalcium phosphate
[4, 5]. (Whitlockite) powders would only be possible in the
Calcium hydroxyapatite (HA) and tricalcium phos- presence of SO2~ anions in precipitation solutions
4
phate (TCP) are the most commonly used calcium containing Ca-nitrate and di-ammonium hydrogen
phosphate ceramics for bone replacement due to their phosphate salts. They reported that all samples which
biocompatibility [3—12]. Variations in atomic Ca/P had not been doped (by about 1% of solution) with
ratios between 2 and 1.5 among the calcium phos- dilute (NH ) SO produced composite ceramics
42 4
phate ceramics, i.e., Ca P O , HA and TCP (b- containing TCP and HA phases, simultaneously.
4 2 9
Whitlockite), are known to lead to differences in Decomposition of the ceramics or fluctuation of the
degradability [13]. It is probably because of this Ca/P ratio is expected to result in such formation of
difference in biodegradation behaviour that biode- unwanted by-products. The objective of this study was
gradable TCP ceramics give rise to extensive bone to synthesize ‘‘pure’’ HA and TCP sub-micrometre
remodelling around the implant [11]. particulated bioceramic powders using aqueous syn-
It has been shown by Asada et al [14] that by- thesis solutions containing dissolved Ca-nitrate
products have been found in HA ceramic powders and di-ammonium hydrogen phosphate salts, and to

0957—4530 ( 1997 Chapman & Hall 91

investigate their high-temperature (800—1300 °C) TA BLE I Experimental conditions for X-ray diffractometer runs
calcination/sintering behaviour and stability in air
atmosphere. Radiation Fe K , 40 kV, 20 mA (incident beam
aÇ
monochromator)
Divergence slit: 0.5°
2. Materials and methods Receiving slit: 0.1 mm
Scattering slit: 0.5°
2.1. Ca-hydroxyapatite ceramic powder Profile breadth: 0.1° 2h
synthesis* Specimen form: Side loaded powder, packed for 2hs
HA powders were prepared by an aqueous precipita- Temperature: 24$2 °C
tion technique using 0.156 M (NH ) HPO (reagent Range of 2h from: 10° 2h to 100° 2h
42 4 Step width: 0.02° 2h
grade, Merck) and 0.4 M Ca(NO ) stock solutions.
32 Counting time: 2 s at each step
Ca(NO ) solutions were prepared by dissolving
32 External 2h standard: Si (Lattice parameter "0.54315 nm)
stoichiometric amounts of CaO (99.9%, BDH Chem- 2h error correction: Linear interpolation from nearest 2hs
icals) in the presence of HNO (reagent grade, Merck). of std.
3 Intensity meas. technique: Automated computer software (Rigaku
The required amounts of both stock solutions to give
the desired Ca/P atomic ratio of 1.667 in the aqueous D-Max/B)
Profile fitting via Pearson VII (Peakfit
mixture were measured, and the pH of both solutions 3.0, Jandel Scientific)
was adjusted to about 10 by the addition of a dilute Cell refinement method: Least-squares (Appleman & Evans,
NH OH solution of proper strength. Following pH 1973).
4
adjustment, 0.096 M (NH ) HPO solution was fed
42 4
from a burette, in a dropwise manner (2—5 ml/min),
thick) in a steel die, without causing a lamination
into a stirred solution of 0.16 M Ca(NO ) heated to
32 problem in the green compacts. The sintering runs
60 °C on a hot plate. After mixing, the solution was
were performed at 1050 ° to 1300 °C (10 h) in air atmo-
aged and vigorously stirred at its boiling point for
sphere. The fired pellets were cooled naturally to the
about 2 h in a sealed container. The HA precipitates
ambient temperature in the furnace.
were recovered from the solution by centrifugal filtra-
Semi-quantitative analyses of the chemical com-
tion and washed three times with distilled water, fol-
position of HA and TCP samples were conducted by
lowed by drying at 100 °C overnight.
EDXS (energy dispersive X-ray spectroscopy, Tracor
Series-2) on a JEOL JSM-6400 SEM. Data acqui-
2.2. Tricalcium phosphate ceramic powder sition was performed at 15 kV for 200 s on the
synthesis gold-coated fracture surfaces of the pellets used for
Stoichiometric proportions of Ca(NO ) and (NH ) morphological analyses. A given particle was analysed
32 42 at the particular magnification at which it filled the
HPO stock solutions were measured and the di-
4 monitor screen. Ca/P mole ratios of the samples were
ammonium hydrogen phosphate aliquot was slowly
added to the Ca-nitrate solution maintained at 60 °C. determined as the average of six consecutive runs on
The pH of the reaction solution was maintained con- each sample. EDXS results are believed to be accurate
stant at 7 by the continuous addition of a dilute to$3%.
NH OH solution. Following a 2 h ageing, the solu-
4
tion was cooled to room temperature and the TCP
precipitates were recovered by centrifugal filtration 3. Results and discussion
from the mother liquor. 3.1. Calcium hydroxyapatite synthesis
Calcium hydroxyapatite precursor powders recovered
from the mother liquor using the procedure outlined
2.3. Phase and microstructure analysis above were observed to display an agglomerate of
The phase purities and crystal structures of the dried extremely fine individual particles of spherical charac-
and lightly calcined (400—600 °C) HA and TCP pow- ter, following drying at 100 °C. This was in agreement
ders were investigated by powder X-ray diffraction with previous observations of Asada et al. [14] and
using a Rigaku D-Max/B diffractometer. Details of Jarcho et al. [15]. Powders calcined at higher temper-
the XRD runs are given in Table I. The powder XRD atures (500—1000 °C) were found to contain 0.6 to
data collected (with counting time increased to 10 s) 0.7 lm spherical and relatively monodispersed par-
on the synthesized HA powders were analysed using ticles. The HA powders had BET specific surface areas
the Rietveld method [17, 22—24] to obtain informa- in the range 45—65 m2/g, the variation being depen-
tion on the HA unit cell contents. dent on the extent of agglomeration present in par-
The morphological features of the bioceramic sam- ticular powder samples. XRD runs indicated that HA
ples were studied through gold-coated fracture surfa- precipitates calcined at 500 °C in air for 6 h pos-
ces of sintered pellets on a scanning electron micro- sessed a hexagonal crystal structure with lattice para-
scope (SEM: JEOL-JSM 6400) at an accelerating volt- meters (least-squares refined with the PC subroutine
age of 15 kV. Particle size was assessed from the SEM Appleman and Evans [18]); a"0.9417 nm and
micrographs. The powders were first lightly ground in c"0.6880 nm. The powder XRD patterns of our HA
an agate mortar and then uniaxially pressed at 25 to samples agreed closely with the JCPDS (Joint Com-
390 MPa into green pellets (1 cm diameter and 3 mm mittee on Powder Diffraction Standards [19]) K 9-
432 file for calcium hydroxyapatite. The XRD data of
* Patent pending, Turkish Patent Institute, Ankara, Turkey. the HA precipitates were analysed in 1 °C 2h intervals
92
(over the range 10 to 100 ° 2h) for more accurate 3.3. HA/TCP composite powder synthesis
detection of weak reflections and possible peak over- The production of composite, sub-micrometre pow-
laps using a commercial spectroscopic profile fitting ders, with a predetermined crystalline HA/TCP ratio
program (Peakfit 3.0, Jandel Scientific, Corte Madera, in the recovered precipitates, was also shown to be
CA, USA) having built-in pseudo-Voigt and Pearson possible, at least for the 40% TCP and 60% HA com-
VII distribution functions. In contrast to some pre- position, with the starting materials and general pro-
vious studies [12, 14, 16, 20] utilizing relatively similar cedures of this study. Work is in progress towards
starting materials and precipitation procedures in the a more precise and systematic determination of the
synthesis of HA powders, our samples were found to experimental conditions of in situ, composite bio-
be free of unwanted by-products or phases such as ceramic powder synthesis.
CaO, a- or b-TCP, Ca P O , CaHPO and Ca P O , The composite powders of 40% TCP—60% HA
2 2 7 4 4 2 9
even after calcining at temperatures as high as 1300 °C composition were synthesized from solutions contain-
in air for 6—10 h. ing calcium nitrate and di-ammonium hydrogen phos-
The process yields (percentage of HA recovered in phate at a Ca/P ratio of 1.60. The solutions were aged
the form of solid precipitates, after calcination at for 2 h at 60 °C and then boiled for 1 h, following
600 °C, from aqueous solutions) during these HA syn- completion of the addition of phosphate solution to
thesis experiments remained constant at around 75%, the nitrate solution. The pH of the reaction solutions
being fairly independent of the bath temperatures (in was kept constant at about 6 by the addition of diluted
the range 25—85 °C) and ageing times (0.5—4 h) studied. HNO or NH OH. The precipitates recovered by
3 4
The boiling step, following completion of the addition centrifugation were later calcined at 500 °C, basically
of di-ammonium phosphate solution into the calcium to ‘‘dry out’’ the samples.
nitrate solution, was found to be quite effective in
decreasing the pH of the solution from 8.5 to 5—5.5,
which is probably accompanied by an advance of the
re-precipitation reaction [11] for the elimination of 3.4. Sintering studies
remnants of the more stable TCP phase The pure HA, TCP, and composite (40% TCP—60%
HA) powders synthesized were first homogenized in
4Ca (PO ) #2H OPCa (PO ) (OH)
3 42 2 10 46 2 an agate mortar and then uniaxially pressed at a pres-
#2Ca2`#2HPO2~ (2) sure range of 180—390 MPa into small pellets. The
4 green pellets prepared at the maximum pressure of
and thus making it possible to obtain pure HA phase 390 MPa (of all compositions) were sintered to above
by the end of the precipitation processes. The re- 99% of the theoretical density at 1200 °C in air in
covered HA precipitates after centrifugation were about 10 h. The sintered samples did not show any
found to be mixed with around 10% TCP phase, espe- signs of structural decomposition or degradation. The
cially at the lower bath temperatures (25—45 °C) main- experimental densities of the sintered pellets were
tained during the process. Increase of bath temper- measured using Archimedes’ liquid-displacement
atures to about 60 °C eliminated the containment method in xylene as the liquid medium. Although
problem. work is still in progress towards full determination of
It was also observed that the delicate washing steps the sintering behaviour of such high surface area and
(with distilled water) following decantation of the high reactivity bioceramic powders, it is quite evident
mother liquor were crucial, to the point that the re- now that with these powders it would not be necessary
tainment of Ca2` and HPO2~ ions in an amorphous to go to temperatures as high as 1400 °C [21] to reach
4
layer surrounding the particles would in turn help full densities, thus avoiding the problems of dehydra-
reaction (2) to take place in the reverse direction at tion and structural decomposition especially for the
calcination temperatures above 800 °C and, thereby, HA powders. Fig. 1 shows the XRD traces of HA,
might explain the consequent observation of TCP, and 40% TCP—60% HA samples sintered at
unwanted phases in synthetic HA powders at high 1200 °C. Since the traces depicted in this figure were
temperatures [12, 14—16, 20]. taken from the direct, as-is surfaces of the sintered
pellets, the peaks were seen to have somewhat broken
tops. The specific crystal reflections for each phase in
3.2. Tricalcium phosphate synthesis the 2h region of interest are labelled as particular
Ca (PO ) was shown to be synthesized more readily planes in each figure. The EDXS analyses performed
3 42
by the procedure described above than HA powders, on these samples confirmed the expected Ca/P ratios
owing to the fact that it did not require an additional of 1.67 (HA), 1.50 (TCP), and 1.60 (40% TCP—60%
boiling step. The precipitates recovered from the solu- HA composite), within the accuracy limits of this
tions were already crystalline and sub-micrometre, as method.
was the case with HA precipitates, and the refined The sintering behaviour of the synthesized sub-mi-
[18] lattice parameters of the hexagonal (of rhom- crometre, spherical HA powders changes, as expected,
bohedral symmetry) TCP phase were measured to be with variations in the initial, green state packing pres-
a"1.0434 nm and c"3.7409 nm, following light cal- sure applied. Fig. 2 shows SEM micrographs of the
cination at 500 °C in air. These parameters were found fracture surfaces of HA pellets pressed uniaxially at
to be in close agreement with the JCPDS PDF K 9- a pressure of 180 MPa as a function of sintering tem-
169 for b-Whitlockite [19]. perature. The pellets of 1 cm diameter and 3 mm
93
 Figure 2 SEM micrographs of HA powders sintered at various
temperatures in air for 12 h after being uniaxially pressed at
180 MPa: (a) 1050 °C; (b) 1200 °C; (c) 1300 °C.

Figure 1 Powder XRD traces of samples sintered at 1200 °C in air

for 10 h after being uniaxially pressed at 390 MPa. (a) calcium
hydroxyapatite, (b) tricalcium phosphate, (c) 40% TCP-60% HA apatitic structure (such as a possible decrease in the
composite powder. intensities of 0 0 4 reflections in the case of dehy-
droxylation problems), according to the powder XRD
results.
thickness were pressed in the presence of 1.5 ml of
2 wt % stearic acid—ethanol solution per gram of pow-
der as a binder. It is apparent from Fig. 2 that the 3.5. Rietveld refinements for
extent of densification abruptly increases with increase crystallographic analyses
in sintering temperature from 1200 °C to 1300 °C (at The Rietveld method [22—24] for the detection and
a constant sintering time of 12 h). Therefore, it could analysis of small structural variations and distortions,
be stated that while the HA pellets pressed at which may not readily be tangible from the bare
180 MPa could only reach a percentage theoretical inspection of the powder XRD traces, is a widely
density of about 97% (Fig. 2c) at 1300 °C after 12 h, accepted tool in the field of inorganic materials and
the pellets prepared from the same powders at ceramics research. The Rietveld method was applied
390 MPa displayed a density of 99.2% at 1200 °C in to the powder XRD data of the HA powder samples of
a reduced time of 10 h. The HA pellets sintered at this study sintered at 1200 °C for 12 h in air. The initial
1300 °C exhibited the crystal structure of pure calcium crystal structure model used in the Rietveld refine-
hydroxyapatite with no signs of decomposition of the ments of these samples was adopted from the single
94
TA BLE I I Positional (unit cell) parameters of the HA powders
(1200 °C)

Atom x y z Site occupancy

O(1) 0.3351(9) 0.4890(9) 0.2500 1.006(29)

O(2) 0.5875(8) 0.4606(9) 0.2500 1.009(42)
O(3) 0.3304(9) 0.2468(11) 0.0687(13) 1.017(79)
P 0.3982(6) 0.3696(6) 0.2500 1.002(4)
Ca(1) 0.3333(2) 0.6667(1) 0.0060(10) 0.6905(12)
Ca(2) 0.2473(6) 0.9914(8) 0.2500 0.9982(23)
O(H) 0 0 0.1767(5) 0.5067(121)
H 0 0 0.0673(5) 0.5032(129)

Estimated standard deviations (e.s.d) for the parameters varied are

given in parentheses following the digits for which they are applic-
able.

crystal structural studies of Young [25, 26] on calcium

hydroxyaptite.
The Rietveld refinements were performed using PC
software [17] on a 486DX2-66 machine. The back-
ground, mixing, half-width, cell, preferred orientation
and peak asymmetry, atomic position, site occupancy
parameters and the anisotropic temperature factors
were refined in each cycle, and the structure factors
were calculated. The number of parameters refined in
the cycles was 64, and the program-generated final
structural model converged successfully within 12
cycles. The weighted R-factor was observed to be
6.9% and the final positional parameters are given in
Table II. The parameters found in this study of the HA
crystal structure were observed to be in close agree-
ment with those reported for a ‘‘single crystal’’ of
natural calcium hydroxyapatite by Young et al. [25].
Examination of the site occupancy factors for syn-
thetic HA revealed that the sum of the values of
calcium atoms (Ca(1) and Ca(2)) found at two crystal-
lographically different sites equals 1.6887. The ratio of
this number to that of the occupancy factor of
P (1.002) would give the mean Ca/P ratio in this
sample as 1.685 $ 0.0102. This means that the HA
samples sintered at 1200 °C do not have any calcium
deficiency and they display a Ca/P ratio very close to
the stoichiometric phase. The low value, 6.9%, for the
weighted R-factor of the Rietveld refinements also
shows that the HA samples do not have any signifi-
cant structural variations or distortions present even
after sintering in air at 1200 °C for about 12 h.
The hexagonal unit cell parameters of the above
sample were found to be a"0.94147(9) nm and
c"0.68777(7) nm by Rietveld analysis, and these
values coincide strongly with the ‘‘single crystal’’ para-
meters reported by Young et al. [25]. Figure 3 Rietveld plots of an HA sample sintered at 1200 ° in air for
One of the typical plots [27] obtained in Rietveld 12 h.
refinements of the powder data on the 1200 °C HA
sample is reproduced in Fig. 3, divided into three 2h
intervals to facilitate easier assessment. This plot de- agreement between the structural model and the raw
picts the extent of disagreement between the structural data.
model fitted and the raw powder XRD data. The dots The crystallite size of the HA powder samples sin-
in Fig. 3 represent the raw data, the solid lines follow- tered at 1200 °C in air was determined to be
ing the dots denote the final structural model, and the 102$2 nm by using the peak half-widths from the
tiny vertical lines at the bottom indicate the peak XRD data according to the Warren—Averbach
positions. The final Rietveld plot displays reasonable method [28, 29]. In calculating the crystallite size of
95
 References
1. M. D . K . BR EM NE R, ‘‘¹he story of dentistry from the dawn of
civilization to the present’’ (Dental Items of Interest Publishing
Company, Brooklyn, NY, 1939) p. 69.
2. H. D R EE SM AN , Beitr. Z. Klin. Chir. ¹uebing 9 (1892) 804.
3. F. H . A L BEE and H . F. MO R R I S ON , Ann. Surg. 71 (1920) 32.
4. T. JA NIK O W S K I and T . D. M C G EE , Proc. Iowa Acad.
Sciences 76 (1969) 113.
5. T. D . M C GE E and J . L . W O O D, J. Biomed. Mater. Res. Symp.
5 (1974) 137.
6. Y . SH I N TO , A. U C H ID A , F. K O RK U S U Z, N . AR AK I and
K. O N O, J. Bone Joint Surg. 74-B (1992) 600.
7. F. K O R K U S U Z, A. U C H I D A, Y . S H IN T O, N . AR AK I ,
K. I N OU E and K. O N O, ibid. 75-B (1993) 111.
8. F. K O RK U S U Z, H . Y E TK I N , K. K A RA M ET E,
B. I RF A N OG L U and N . A K K AS , Engineering Systems
Design and Analysis, ASME 1994; 64-4: 83.
9. C . LA VE RN I A and J. M. SC HO E NU N G, Ceram. Bull. 70
Figure 4 The Warren—Averbach plot of an HA sample sintered at (1991) 95.
1200 °C in air for 12 h for the determination of crystallite size. 10. G. G. NI ED ER AU E R, T. D . M C G EE and R . K. K U D EJ,
ibid. 70 (1991) 1010.
11. C . P . A . T. K L EIN , J. M. A. D E BLI EC K -H O G ER VO R S T,
J. G . C . W O LK E and K . D E G R O O T, Biomaterials 11 (1990)
509.
the samples, diffraction angles corresponding to the 12. I. O R L Y , M . G RE G O RIE, J. M E NA N TE AU , M. H E U -
GH E BA ERT and B. KE R EBE L, Calcif. ¹iss. Int. 45 (1989) 20.
peak positions and the full width at half maximum
13. C . P. A. T . K LEI N , A. A . D RI ES S E N, K . D E G R O O T and
values were used. The numerical values of the terms A. V A N D EN H OO F F, J. Biomed. Mater. Res. 17 (1983) 769.
defining the x- and y-axes of Fig. 4 were calculated, 14. M. A SAD A , Y . MI U R A, A. O S AK A , K . O U K AM I and S.
and a linear regression was performed on the gener- NA KA MU RA, J. Mater. Sci. 23 (1988) 3202.
ated data points. The intercept of the regression 15. M. J A RC H O , C. H . BO L EN , M . B . TH O M AS, J . B O BI CK ,
J. F . KA Y and R . H. D OR EM US, ibid. 11 (1976) 2027.
line was found to be at 0.009 635 9 which correspon-
16. M. J AR C HO , R . L. S AL S BU RY , M. B . TH O M AS and R. H .
ded to the square root of the reciprocal of the crystal- DO RE MU S, ibid. 14 (1979) 142.
lite size. 17. A . S A K TH I V EL and R . A. YO U N G, ‘‘Program DBWS-9411
PC’’. Version: March 1995. School of Physics, Georgia Insti-
tute of Technology, Atlanta, GA 30332, USA.
4. Conclusions 18. D. E. A PP L EM A N and H . T . E VA N S, º.S. Geological Sur-
vey, Computer Contribution No. 20, NTIS Report No. PB-
Sub-micrometre, spherical, and chemically homogene- 216188. U.S. National Technical Information Service, Spring-
ous powders of calcium hydroxyapatite (HA), trical- field, VA, USA, 1973.
cium phosphate (TCP), and composite powders 19. Joint Committee on Powder Diffraction Standards, 1600 Park
containing tailorable amounts of both of the above Lane, Swarthmore, PA 19081, USA.
two phases were synthesized by chemical precipitation 20. K. I O KU , M. Y O S H IM UR A and S . S O M IY A , Biomaterials
11 (1990) 57.
from aqueous solutions of nitrate and phosphate salts. 21. K. Y AM AS H I T A, K . K I TA G AK I , T . U M EG A KI and
The sintering and densification behaviours of the pow- T. K A NA Z A W A, J. Mater. Sci. ¸ett. 9 (1990) 4.
ders were monitored. It was seen through crystallo- 22. H. M . R IET V ELD, Acta Crystallogr. 22 (1967) 151.
graphic analyses based on the atomic positions in the 23. idem., J. Appl. Crystallogr. 2 (1969) 65.
unit cells of HA powders that the synthesized powders 24. D. B . W IL ES and R . A. Y O UN G, ibid. 14 (1981) 149.
25. K. SU D AR S A NA N and R. A. YOUNG, Acta Crystallogr. B25
remained pure and uniform in chemical composition, (1969) 1534.
even after sintering at 1300 °C for prolonged times. 26. idem., ibid. B34 (1978) 1401.
The attainment of percentage theoretical densities in 27. H. M AR C I N I AK , R . D I D U SZ KO , ‘‘DMPLOT’’, Version
excess of 99% was easily achieved in cold samples, 3.41beta, Jan. 1995. High Pressure Research Center UNI-
uniaxially pressed for about 10 h at a relatively low PRESS, Sokolowska 29, 01-142 Warsaw, Poland.
28. B. E . W AR RE N, ‘‘Impact crystals, X-ray diffraction,’’ (Ad-
temperature of 1200 °C in air, with no signs of struc- dison-Wesley, Monlo Park, CA, 1969) Chapter 13.
tural decomposition. It was also shown that the degree 29. Y . Z H AN G, J . M . S T EW A RT , B. M O RO S IN and C. R .
of porosity in the pressed samples could be controlled HU B BA RD, in ‘‘Ceramic transactions, Vol. 1, Ceramic Powder
by careful monitoring of the heating schemes. Science IIB’’, edited by G. L. Messing, E. R. Fuller and H.
Hausner (The American Ceramic Society, Westerville, OH,
1988) p. 1192.

Acknowledgements
This research was supported through the project Received 29 September 1994
METU AFP 94-03-08-03. and accepted 14 March 1996

96