Journal J. Am. Ceram. Soc.

, 82 [10] 2813–18 (1999)

Phase Evolution during the Formation of ␣-Tricalcium Phosphate

Kevor S. TenHuisen*,† and Paul W. Brown*
Materials Research Laboratory, The Pennsylvania State University, University Park, Pennsylvania 16802

The formation of ␣-tricalcium phosphate from monobasic calcium hydroxide.8–11 To form phase-pure HAp, TetCP must
ammonium phosphate and calcium carbonate was investi- be combined with another phosphate more acidic than HAp.
gated using a number of complementary techniques. Dif- These may include phosphoric acid, monocalcium phosphate
ferential thermal analysis showed five distinct thermal monohydrate (Ca(H2PO4)2䡠H2O), dicalcium phosphate dihy-
events attributed to melting of the ammonium phosphate, drate (CaHPO4䡠2H2O), dicalcium phosphate (CaHPO4), octa-
decomposition of acidic calcium orthophosphate into an calcium phosphate pentahydrate (Ca8(HPO4)2(PO4)4䡠5H2O),
amorphous calcium metaphosphate, crystallization of and ␣-TCP. The reactivities of these calcium phosphate ce-
␤-calcium metaphosphate, crystallization of ␤-calcium py- ments (CPCs) are functions of the respective particle sizes12 of
rophosphate, and calcination of calcium carbonate. X-ray the precursors and their impurity contents.
diffraction analyses as a function of temperature supported Typically, high-temperature phases, such as SHAp, TetCP,
the evolution of these events and phases and also showed ␣-TCP, and ␤-Ca3(PO4)2 (␤-TCP), are synthesized from cal-
the formation of other intermediates, including an amor- cium carbonate (CaCO3) and either calcium pyrophosphate
phous phase, apatite, lime, and ␤-tricalcium phosphate. (Ca2P2O7) or dicalcium phosphate dihydrate. While phase pu-
Thermal gravimetric analysis (TGA) showed a large weight rity can be achieved by these synthesis routes, overfiring re-
loss occurring between 150° and 250°C due to reaction be- sults in deadburning. This results, in turn, either in low reac-
tween the acidic phosphate liquid and the calcium carbon- tivity materials and/or in materials that require excessive
ate. This resulted in an amorphous intermediate with a grinding, resulting in contamination. SHAp and TCP can also
Ca/P ratio between 0.5 and 1.0 from which both ␤-calcium be formed by heating salt solution precipitates that have the
metaphosphate and ␤-calcium pyrophosphate crystallized. proper stoichiometries, but it is difficult to completely remove
Mass spectroscopy indicated that the ammonia and carbon residual cationic and anionic species, such as sodium, potas-
dioxide were evolved in four different steps, while water sium, and chloride ions. It has been shown that the rate of
was evolved in at least five steps. These decomposition steps hydration of ␣-TCP is significantly affected by the synthesis
correlated with those observed by TGA. Scanning electron route.1 TCP formed from CaCO3 and either Ca2P2O7 or
microscopy indicated the formation of an intermediate CaHPO4䡠2H2O is more reactive than that synthesized from pre-
phase that coated the calcium carbonate by 250°C. The cipitated amorphous calcium phosphate (ACP) or CDHAp. The
proposed mechanistic reaction path to ␣-tricalcium phos- present study investigates the mechanistic path taken when
phate involves the formation and consumption of the fol- monobasic ammonium phosphate (NH4H2PO4) and CaCO3 are
lowing sequence of intermediates: an acidic amorphous pyroprocessed to produce ␣-TCP.
condensed phosphate; ␤-calcium metaphosphate and ␤-
calcium pyrophosphate; lime, apatite, and ␤-tricalcium II. Experimental Procedure
phosphate.
Certified ACS grades of NH4H2PO4 (Fisher Scientific Co.,
I. Introduction Pittsburgh, PA) and CaCO3 (Fisher Scientific) were combined
in a 2/3 molar ratio and milled in anhydrous ethanol for 4 h
T HE formation of calcium phosphates at high temperature is
of particular interest for biomedical, phosphor, and dinner-
ware applications. Optimization of processing reduces the in-
using zirconia media. The slurry was immediately vacuum fil-
tered and dried. The filter cake was stored in a desiccator over
P2O5 until it was used.
troduction of impurities and can be of critical importance for
Thermal gravimetric analysis (TGA) (Model Delta, Series
materials being processed for in vivo and phosphor use. Hy-
TGA7, Perkin-Elmer Corp., Norwalk, CT) and differential
droxyapatite (HAp) formed by high-temperature methods is
thermal analysis (DTA) (Model DTA1700, Series 7/4, Perkin-
used routinely in the replacement of bone tumors and augmen-
Elmer) were performed on ∼50 mg of precursor powder over a
tation procedures. Among the other calcium orthophosphates
temperature range of 25°–1250°C, utilizing a heating rate of
being investigated for in vivo applications are those that hy-
10°C/min. TGA was performed in flowing air (6 cm3/min).
drolyze in aqueous solutions and produce HAp by a cementlike
Fifteen pieces of the precursor cake (∼1.5 g each) were
reaction. These include tetracalcium phosphate (Ca4(PO4)2O,
placed on platinum foil and heated at a rate of 10°C/min in a
TetCP) and ␣-tricalcium phosphate (␣-Ca3(PO4)2, ␣-TCP).
MoSi2 furnace. Samples were removed at 200°, 250°, 300°,
While ␣-TCP reacts directly with water to form a calcium-
400°, 500°, 600°, 700°, 800°, 900°, 1000°, 1100°, 1150°,
deficient HAp (Ca9HPO4(PO4)5OH, CDHAp),1–7 TetCP hy-
1200°, and 1250°C. These samples were immediately ground
drates into stoichiometric HAp (Ca10(PO4)6(OH)2, SHAp) and
in an alumina mortar and pestle, placed into plastic scintillation
vials, and stored in a desiccator over P2O5 under vacuum. Each
sample was packed into a zero background quartz recessed
M. Grutzeck—contributing editor slide. Powder X-ray diffraction (XRD) analysis was performed
using a powder diffractometer (Model PAD V, Scintag Inc.,
Sunnyvale, CA) with CuK␣ radiation, a graphite monochrom-
Manuscript No. 191397. Received November 6, 1996; approved March 16, 1998. eter, and a scintillation detector. Samples were scanned over a
Supported by the National Institutes of Health, National Institute of Dental Re- 2␪ range of 15°–40° using a step size of 0.02° and a scan rate of
search Division under Grant No. DE09421.
Based in part on the dissertation submitted by K. S. TenHuisen for the degree of 2°/min. Selected unground samples were also observed uncoated
Ph.D. in materials at The Pennsylvania State University, University Park, PA.
*Member, American Ceramic Society.
in an environmental scanning electron microscope (ESEM;
†Now at Johnson and Johnson, Johnson and Johnson Corporate Biomaterials Cen- Model Electroscan, FEI Beam Technology, Hillsboro, OR).
ter, Somerville, NJ. In addition to analyzing condensed phase evolution, we ana-
2813
2814 Journal of the American Ceramic Society—TenHuisen and Brown Vol. 82, No. 10

lyzed the evolution of gas during heating. A mass spectrometer tion with temperature as determined by XRD analyses. The
(Model MS-25RFA, Kratos Analytical, Inc., Manchester, NH4H2PO4 is preferentially consumed during the initial stage
U.K.) with a double sector magnet was used. A small amount of the reaction. By 250°C, there is no evidence of this acidic
of precursor powder was placed into a quartz ampule and in- precursor. The only observable crystalline phase is CaCO3.
serted by a direct insertion probe into the mass spectrometer for Bulk stoichiometry, thus, dictates the presence of an acidic
analysis of H2O, CO2, and NH3. Source pressure and tempera- ACP intermediate.
ture were ∼10−6 torr (∼1.3 × 10−4 Pa) and 100°C, respectively. Between 250°–500°C, a continuous loss in weight is ob-
A Kinderman probe was used to heat the sample between 100° served (∼8.75%). Only CaCO3 is observed by XRD at tem-
and 500°C at a rate of 10°C/min. Ionization was induced by peratures <500°C, but its quantity, as estimated by diffraction
electron impact (EI) at ∼70 eV. The magnet was scanned con- intensity, decreases during this time. This suggests that the
tinuously from 130 amu (atomic mass unit) to 16 amu. A amorphous phase(s) present are slowly increasing in their Ca/P
background was obtained prior to the experiment to determine ratio. At 500°C, crystallization of an intermediate calcium
the contributions of atmospheric H2O and CO2 to the data. phosphate from the amorphous phase(s) are observed by XRD
Each of the three gaseous species was integrated at 10°C in- (see Fig. 2).
tervals to obtain their relative cumulative losses as a function of Although only two new diffraction peaks are observed, they
temperature. Because EI results in two species for both NH3 appear to result from the formation of ␤-calcium metaphos-
(16 amu (NH+2) and 17 amu (NH+3)) and H2O (17 amu (OH+) phate (␤-Ca(PO3)2). A previous study13 on the evolution of
and 18 amu (H2O+)), NH3 was determined using an amu of 16 phases with heating of a variety of calcium orthophosphate
and H2O using an amu of 18. phases has shown that the acidic phase, Ca(H2PO4)2䡠2H2O
(MCPM), can follow two different decomposition–phase trans-
formation paths. In the presence of H2O vapor, MCPM decom-
III. Results and Discussion poses to a calcium acid pyrophosphate (CaH2P2O7) at 270°–
280°C, followed by another decomposition into ␥-calcium
(1) Thermal Analysis and Phase Evolution metaphosphate at 340°–360°C. This metaphosphate then trans-
Figure 1 shows the cumulative and rate of weight loss curves forms into other polymorphs, ␤- (450°–500°C) and ␣- (970°C),
during a reaction between NH4H2PO4 and CaCO3 to form as the temperature is further increased. In the absence of steam,
␣-Ca3(PO4)2. The bulk reaction is shown in Eq. (1). MCPM decomposes into an amorphous calcium metaphos-
phate at ∼250°C. This amorphous phase then crystallizes into
2NH4H2PO4 + 3CaCO3 → ␣-Ca3(PO4)2 ␤-Ca(PO 3 ) 2 around 450°–500°C. The appearance of
+ 2NH3↑ + 3CO2↑ + 3H2O↑ (1) ␤-Ca(PO3)2 by 500°C indicates a portion of the amorphous
The theoretical total weight loss due to evolution of all gas- intermediate is composed of long chains of condensed phos-
eous species is 41.51%. Calculated weight losses resulting phate of the general form (PO3)n. (It has been shown that the
from the individual gases (NH3, CO2, and H2O) are 6.42%, crystal structure of this phase consists of “meandering chains”
24.90%, and 10.19%, respectively. The total weight lost mea- of PO4 tetrahedra.14) Because an amorphous intermediate and
sured upon heating to 1250° is 41.48%; this compares well ␤-Ca(PO3)2 are the only phases observed, the H2O vapor
to the theoretical value. TGA shows all volatiles are removed evolved during the reaction must not be sufficient to induce the
by 1000°C. Above this temperature, no change in weight is initial transformation path discussed.
observed. The phase evolution just discussed is supported by DTA
The weight loss between ∼150° and ∼250°C (16.1%) results results from this reaction. A large endotherm (labeled event 1
from the initial reaction between NH4H2PO4 and CaCO3. in Fig. 1) beginning just before 200°C is a result of the melting
NH4H2PO4 melts around 190°C and forms an acidic liquid. of NH4H2PO4. While a reaction between an acidic liquid and
The acidic liquid then reacts with the CaCO3, evolving CO2. CaCO3 should result in an exotherm, the endothermic effects of
However, it is shown that all three gases are being evolved gas evolution and the melting process dominate the DTA re-
during this initial reaction stage. Figure 2 shows phase evolu- sults in this experiment and result in the absence of an exo-
therm. The second DTA thermal event (event 2) shown in Fig.
1 is also an endotherm and begins around 250°C. This event
occurs near the temperature reported for the decomposition of
MCPM into an amorphous calcium metaphosphate. Although
no evidence of crystalline MCPM is observed, the Ca/P ratio of
the intermediate(s) is small and is shown by mass spectroscopy
to be <1 at this temperature. It is likely that the rapid changes
occurring between 200°–250°C do not allow the formation of
crystalline intermediates.
The third DTA thermal event (event 3 in Fig. 1) is a small
exotherm occurring around 450°C. This is near the temperature
at which amorphous calcium metaphosphate has been shown to
transform into ␤-Ca(PO3)2.13 This is supported by XRD analy-
ses in that indications of a crystalline intermediate only become
apparent by 500°C.
Between 500°–750°C, the ␤-Ca(PO3)2 increases in quantity,
as indicated by XRD peak intensities. Over this temperature
range, an additional 2.3% of the weight is lost. At 700°C, the
only significant phases observed by XRD are still CaCO3 and
␤-Ca(PO3)2, but there are small indications of the formation of
␤-Ca2P2O7. The formation of ␤-Ca2P2O7 by this temperature
would support DTA and TGA results showing a small exo-
therm (event 4 in Fig. 1) and increased rate in weight loss
(event 4 in Fig. 1) between 550°–575°C. This could be due to
Fig. 1. Thermal analysis data for the decomposition reaction of the crystallization of a small amount of amorphous Ca2P2O7䡠
NH4H2PO4 and CaCO3 to ␣-TCP at a heating rate of 10°C/min. Curve 4H2O.13
(a) is the TGA cumulative weight loss; curve (b), TGA rate of weight Between 750°–900°C, another rapid loss in weight is ob-
loss; curve (c), DTA. served (event 5 in Fig. 1). This results from the decomposition
October 1999 Phase Evolution during the Formation of ␣-Tricalcium Phosphate 2815

Fig. 2. XRD patterns showing the complex phase evolution with temperature for reaction between NH4H2PO4 and CaCO3 to ␣-TCP. Furnace ramp
rate was 10°C/min. Temperatures at which samples were removed are shown. Labels are as follows: AP (NH4H2PO4), CC (CaCO3), MP
(␤-Ca(PO3)2), PP (␤-Ca2P2O7), A (Ca10(PO4)6(OH)2), L (CaO).

of the remaining CaCO3 into CaO and corresponds to the final at 800°C. By 900°C, the ␤-Ca(PO3)2, HAp, and CaCO3 have
DTA endotherm shown in Fig. 1 (event 5). Over this tempera- been consumed. ␤-TCP appears by 900°C with CaO and
ture range, an additional 14.35% weight loss is observed. If this ␤-Ca2P2O7 still present. At 1100°C, XRD indicates phase-pure
weight loss is attributed solely to the decomposition of the ␤-TCP.
remaining CaCO3, then 42% of the original CaCO3 is con- By 1150°C, a mixture of both ␣- and ␤-TCP are present. The
sumed prior to its calcination temperature and results from its ␤-TCP has completely transformed by 1250°C. The tempera-
reaction with the acidic precursor and intermediate(s). This ture reported for the polymorphic transition of ␤- to ␣-TCP
indicates that the bulk Ca/P molar ratio of the intermediate varies greatly with reported values ranging between ∼1125°–
phases is ∼0.63 at 750°C. Assuming the existence of only two 1400°C.15,16 This variability in the transformation temperature
intermediate phases at this temperature results in a 75/25 molar is a result of the stabilizing effect of magnesium on the ␤-TCP
mixture of Ca(PO3)2/Ca2P2O7. This analysis is consistent with polymorph. It has been shown that if one magnesium atom
the presence of both ␤-Ca2P2O7 and ␤-Ca(PO3)2 at 700°C. substitutes for every 30 calcium ions, the transformation tem-
The calcination of CaCO3 is accompanied by the formation perature is elevated to 1450°C.17 Therefore, to minimize the
of other intermediates. XRD indicates the presence of temperature and firing time of ␣-TCP synthesis, the magne-
␤-Ca2P2O7, HAp, and CaO, along with ␤-Ca(PO3)2 and CaCO3 sium content must be kept at a minimum. Because MgCO3 is
2816 Journal of the American Ceramic Society—TenHuisen and Brown Vol. 82, No. 10

a common compound associated with CaCO3, the selection of resulting from the evolution of gas are observed at this tem-
the carbonate source is critical. The CaCO3 initially used in this perature. The largest change in sample dimension is also ob-
study contains 0.49 wt% magnesium as MgO. When combined served at this temperature, which correlates to the TGA data
with NH4H2PO4 and fired, a mixture of ␤-TCP and ␣-TCP is indicating that ∼40% of the total weight lost by this reaction
observed at firing temperatures and times >1300°C. When has occurred by 250°C. The greatest likelihood of expansion
high-purity CaCO3 containing 0.01 wt% magnesium as MgO is occurs when the NH4H2PO4 initially melts and there is a liquid
used, phase-pure ␣-TCP is produced at <1250°C, as described. phase present while gases are rapidly being evolved. As the
If ramped to temperature and held for 0.5–1 h, phase-pure intermediate phase becomes richer in calcium, it transforms
␣-TCP can be synthesized at 1125°–1150°C using this high- from a liquid to a solid phase.
purity CaCO3 precursor with NH4H2PO4. As the temperature increases, the amorphous intermediate
phase continues to react with the CaCO3, and larger intercon-
(2) Microstructural Evolution nected pores become more apparent (see Fig. 3(c)). Between
Figure 3 shows the changes that occur in microstructure 300°–800°C, significant microstructural changes are not ob-
upon heating NH4H2PO4 and CaCO3 to produce ␣-TCP. These served, even though TGA analysis indicates ∼30% of the total
microstructures are those resulting from the firing study used weight lost by this reaction occurs over this temperature range.
for XRD analyses. Figure 3(a) shows the microstructure of the This likely occurs because of the solidification of the amor-
precursor material before heating. The precursors are relatively phous intermediate phase as its Ca/P ratio increases. The ap-
homogeneous and characterized by particles less than ∼15 ␮m. pearance of coated CaCO3 particles and shells becomes appar-
Light scattering particle size analysis of the precursors after ent over this temperature range as the CaCO3 is slowly consumed.
4 h of milling indicate that the mean size of the CaCO3 is ∼8 ␮m. Significant microstructural changes are again observed at
Large changes in microstructure occur upon heating to 900°C (see Fig. 3(d)). By this temperature, TGA shows that the
250°C (see Fig. 3(b)). XRD analysis of this sample indicates majority of the gases from this reaction have been evolved.
the presence of only CaCO3 and an amorphous phase. This XRD indicates the presence of numerous crystalline interme-
amorphous phase is apparent in Fig. 3(b), appearing as a coat- diate phases (see Fig. 2). A complex pore structure containing
ing that conceals any indications of particulate CaCO3. Pores macroporosity and microporosity is again apparent and is a

Fig. 3. Microstructure of homogenized ␣-TCP precursors and the reaction products resulting from heating at various temperatures during ␣-TCP
synthesis. (a) ␣-TCP precursors (NH4H2PO4 and CaCO3) prior to heating (bar ⳱ 10 ␮m). (b) Reaction product after heating to 250°C. Note the
appearance of a melted phase containing pores (bar ⳱ 10 ␮m). (c) Reaction product after heating to 600°C. This microstructure is representative
of that observed between 300°–800°C (bar ⳱ 10 ␮m). (d) Reaction product after heating to 900°C (bar ⳱ 10 ␮m). (e) Reaction product after heating
to 1150°C. Note the disappearance of the phase(s) that coated the CaCO3 at lower temperatures (bar ⳱ 10 ␮m). (f ) Reaction product after heating
to 1150°C. Note the pseudomorph of a CaCO3 crystal (bar ⳱ 2 ␮m). (g) Reaction product after heating to 1200°C (bar ⳱ 10 ␮m). (h) Reaction
product after heating to 1200°C. Note the high degree of macroporosity and microporosity (bar ⳱ 50 ␮m).
October 1999 Phase Evolution during the Formation of ␣-Tricalcium Phosphate 2817

result of the calcination of the remaining CaCO3, resulting in

CaO formation and CO2 evolution. Although there are pseu-
domorphs of the prismatic CaCO3 morphology (top middle and
top right of Fig. 3(d)), XRD indicates its absence by this tem-
perature. The core of this microstructure is likely the unreacted
CaO observed by XRD analyses.
Figures 3(e) and 3(f) show representative microstructures
after heating to 1150°C. XRD predominantly shows the two
polymorphs of TCP. CaCO3 ghosts are still present, and the
microstructure appears more “particulate” in nature. The con-
sumption of the CaCO3 “shell” and its CaO core is shown in
Fig. 3(f). By 1200°C, the morphology reminiscent of CaCO3 is
no longer observed, and necking and localized sintering of the
␣-TCP particles become apparent (see Figs. 3(g) and (h)).
(3) Mass Spectroscopy
The relative rates of NH3, H2O, and CO2 evolution with
temperature are shown in Fig. 4. The maximum rates for each
species is shown in ion counts. Both NH3 and CO2 appear to
leave in four steps, while the H2O leaves in at least five steps.
Adjusting the absolute rates of each species (because of frag-
mentation) to represent the total amount of each gas being
evolved and then summing as a function of temperature results
in the curves shown in Fig. 5. These cumulative losses are
Fig. 5. Relative cumulative losses of NH3, H2O, and CO2 with tem-
plotted on a molar basis. From the beginning of heating, it can perature for reaction between CaCO3 and NH4H2PO4 to form ␣-TCP.
be seen that all three gases are rapidly being evolved, support-
ing previous observations about the large volume change in the
sample between 200°–300°C. Although CaCO3 does not de-
compose until 850°–900°C, the acidic liquid formed as
NH4H2PO4 melts results in an initial rapid acid–based reaction
between the two precursors and results in the evolution of all
three gaseous species.
Figure 6 shows the cumulative loss of all three gases based
on weight. This figure is a thermal gravimetric curve con-
structed from the results of mass spectroscopy. Comparing Fig.
1 to Fig. 6 shows the effects of measuring weight losses at
different pressures. Because mass spectroscopy is performed
under high vacuum, thermal events occur at lower temperatures
when compared to TGA performed at atmospheric pressure.
The initial maximum rate in weight loss occurs around 160°C
for mass spectroscopy (event 1 in Fig. 6) as compared to 210°C

Fig. 6. Relative TGA curves (rate and cumulative) constructed from

mass spectroscopy data on NH3, H2O, and CO2 evolution during re-
action between CaCO3 and NH4H2PO4 to form ␣-TCP.

for the TGA experiment (event 1 in Fig. 1). As the temperature

is increased, the difference in temperature between the same
thermal events slightly increases. The two features labeled
event 3 and event 4 in the TGA plot (Fig. 1) correspond to
those events containing the same labels in Fig. 6. Both events
3 and 4 are shifted by ∼90°C. Thus, the reaction at the end of
the mass spectroscopy experiment (500°C) is probably closer
to that occurring under atmospheric pressures at 600°C.
Comparing Figs. 4–6 shows event 1 results from the evolu-
tion of all three gases as the NH4H2PO4 melts and reacts with
the CO3. Event 3 also involves the evolution of all three gases
with the greatest contribution from H2O. This implies that
Fig. 4. Chromatograms showing the rate of NH3 (16 amu), H2O (18 small amounts of NH3 are still present in the system up to
amu), and CO2 (44 amu) evolution with temperature. Maxima for 450°–500°C at atmospheric pressure. This is possible because
NH3, H2O, and CO2 correspond to ion counts of 4.31 × 106, 3.04 × a number of different ammonium hydrogen and calcium am-
106, and 6.28 × 106, respectively. monium hydrogen phosphates have been reported to exist.18
2818 Journal of the American Ceramic Society—TenHuisen and Brown Vol. 82, No. 10

The fourth event appears to be a result of only H2O evolu- NH3/H2O ratio, which shows a decrease in the acidity (i.e., H+)
tion and supports the previous discussion attributing this event of the phosphate intermediate.
to the crystallization of amorphous Ca 2 P 2 O 7 䡠4H 2 O into
␤-Ca2P2O7. IV. Conclusions
Figure 7 plots the cumulative molar ratios of NH3/H2O and
CO2/NH3 evolved as a function of temperature. The molar ratio Reaction between NH4H2PO4 and CaCO3 mixed in a 2/3
of NH3/H2O reflects the degree to which the phosphate has molar ratio followed a complex reaction path to form ␣-TCP.
condensed through the course of the reaction. This relationship Initially, the NH4H2PO4 melted and formed an acidic liquid
can be seen in Eq. (2). As the NH3/H2O ratio decreases, the that reacted with a portion of the CaCO3 between 200°–250°C.
value of x in Eq. (2) becomes larger, and the degree of phos- During this reaction, large volume changes occurred in the
phate condensation increases. In the absence of a second phase sample because of the evolution of H2O, NH3, and CO2 and the
(CaCO3), complete condensation of phosphate and evolution of presence of a liquid phase. The NH4H2PO4 was observed to
all available NH3 and H2O would result in a NH3/H2O ratio of disappear by 250°C, resulting in only one crystalline phase,
0.66. This ratio shows a general decrease with increasing tem- CaCO3, being observed between 250°–500°C. The intermedi-
perature, inferring continuous phosphate condensation below ate amorphous phase observed between these temperatures was
600°C. shown to have a Ca/P ratio between 0.5 and 1.0, calculated to
be ∼0.63 at 700°C by TGA. From this amorphous intermediate
NH4H2PO4 → NH3 + xH2O + (H3−2xPO4−x)intermediate (2) phase, ␤-Ca(PO3)2 initially crystallized around 450°–500°C
followed by ␤-Ca2P2O7. Between 800°–900°C, the remaining
The CO2/NH3 ratio reflects changes in the Ca/P ratio of the CaCO3 decomposed with the formation of CaO, HAp, and
intermediate phase(s). NH4H2PO4 is shown by XRD to be con- increasing amounts of ␤-Ca2P2O7. By 1100°C, XRD showed
sumed by 250°C, and the majority of the NH3 has been evolved phase-pure ␤-TCP, which transformed into the ␣-phase by
by this temperature, as shown by mass spectroscopy. This in- ∼1200°C.
dicates the total phosphate present in the intermediate should Although the present study explored the formation of
be equivalent to the amount of NH3 evolved, based on the 1/1 ␣-TCP, the approach can be used to make a variety of calcium
molar ratio as shown in Eq. (2). phosphates, including ␤-TCP, HAp, and TetCP. In addition,
The proportion of calcium present in the intermediate should the multimethod analyses described are applicable for optimiz-
also be equivalent to the amount of CO2 evolved because of the ing the pyroprocessing steps in producing a variety of ceramics
reaction shown in Eq. (3). by understanding more fully the reaction paths occurring dur-
ing heating and in lowering the levels of contamination by
CaCO3 → CO2↑ + (CaO)intermediate (3) minimizing the extent of subsequent comminution that may be
Although the ratios calculated by this technique are not exact, required.
approximate values and trends in the data can be obtained by
analyzing them. As shown in Fig. 7, the CO2/NH3 ratio is less References
1
than unity (between 0.65–0.85), consistent with XRD and TGA H. Monma, M. Goto, and T. Kohmura, “Effect of Additives on Hydration
and Hardening of Tricalcium Phosphate,” Gypsum & Lime, 188, 11–16 (1984).
findings showing the Ca/P ratio in the intermediate prior to the 2
H. Monma and T. Kanazawa, “The Hydration of ␣-Tricalcium Phosphate,”
calcination of the remaining CaCO3 is between 0.5 and 1.0 Yogyo Kyokaishi, 84, 209–13 (1976).
3
(0.63). The CO2 /NH3 ratio generally increases with tempera- H. Monma and T. Kanazawa, “Wet-Process Formation of Nonstoichiometric
ture, as would be expected as the CaCO3 slowly reacts with the Hydroxyapatite from Tricalcium Phosphate,” Yogyo Kyokaishi, 86, 73–76
(1978).
phosphate-rich intermediate phase. The increase in the Ca / P 4
H. Monma, S. Ueno, and T. Kanazawa, “Properties of Hydroxyapatite Pre-
ratio with temperature is also reflected in the decrease in the pared by the Hydrolysis of Tricalcium Phosphate,” J. Chem. Technol. Biotech-
nol., 31, 15–24 (1981).
5
H. Monma, S. Ueno, M. Tsutsumi, and T. Kanazawa, “Changes in Compo-
sition and Structure of Tricalcium Phosphate under Different pH Conditions,”
Yogyo Kyokaishi, 86, 590–97 (1978).
6
K. S. TenHuisen and P. W. Brown, “Formation of Calcium-Deficient Hy-
droxyapatite from ␣-Ca3(PO4)2,” Biomaterials, 19, 2209–17 (1998).
7
K. S. TenHuisen and P. W. Brown, “The Effects of Homoepitaxy on the
Formation of Calcium-Deficient Hydroxyapatite from ␣-Ca3(PO4)2,” J. Biomed.
Mater. Res., in review.
8
W. E. Brown and L. C. Chow, “A New Calcium Phosphate Water Setting
Cement”; pp. 352–77 in Cements Research Progress—1986, Edited by P. W.
Brown. American Ceramic Society, Westerville, OH, 1987.
9
R. I. Martin and P. W. Brown, “Hydration of Tetracalcium Phosphate,” Adv.
Cem. Res., 5, 119–25 (1993).
10
L. Xie and E. A. Monroe, “The Hydrolysis of Tetracalcium Phosphate and
Other Calcium Orthophosphates”; pp. 29–37 in Handbook of Bioactive Ceram-
ics, Volume II, Calcium Phosphate and Hydroxylapatite Ceramics. Edited by T.
Yamamur, L. L. Hench, and J. Wilson. CRC Press, Boca Raton, FL, 1990.
11
L. Xie and E. A. Monroe, “Calcium Phosphate Dental Cements”; pp. 25–39
in Specialty Cements with Advanced Properties, Materials Research Society
Symposium Proceedings, Vol. 179. Edited by B. Sheetz, A. Landers, I. Odler,
and H. Jennings. Materials Research Society, Pittsburgh, PA, 1991.
12
L. C. Chow, S. Takagi, P. D. Constantino, and C. D. Friedman, “Self-
Setting Calcium Phosphate Cements”; see Ref. 11, pp. 3–24.
13
A. O. McIntosh and W. L. Jablonski, “X-ray Diffraction Powder Patterns of
the Calcium Phosphates,” Anal. Chem., 28, 1424–27 (1956).
14
W. Rothammel, H. Burzlaff, and R. Specht, “Structure of Calcium Meta-
phosphate Ca(PO3)2,” Acta Crystallogr., Sect. C: Cryst. Struct. Commun., C45,
551–53 (1989).
15
E. M. Levin, C. R. Robbins, and H. F. McMurdie; Fig. 246 in Phase Dia-
grams for Ceramists. Edited by M. K. Reser. American Ceramic Society, Co-
lumbus, OH, 1964.
16
J. H. Welch and W. Gutt, “High-Temperature Studies of the System Cal-
cium Oxide–Phosphorus Pentoxide,” J. Chem. Soc., London, 4442–44 (1961).
Fig. 7. Ratio of gaseous species evolved as a function of temperature 17
J. Ando, “Tricalcium Phosphate and its Variation,” Bull. Chem. Soc. Jpn.,
showing the approximate Ca/P molar ratio (by CO2/NH3) in the in- 31, 196–201 (1958).
termediate(s) and the degree of condensation of the phosphate groups 18
A. Durif, Crystal Chemistry of Condensed Phosphates; pp. 15–186. Ple-
(by NH3/H2O). num, New York, 1995. 䊐