PROD.
#:1987 3739-1 - JBMR - Wiley - 3739N1 / c104-15 / 01-24-97 13:19:59
Effect of atmosphere on phase transformation in plasma-
sprayed hydroxyapatite coatings during heat treatment
Jiyong Chen,* Weidong Tong, Yang Cao, Jiaming Feng, and Xingdong Zhang
Institute of Materials Science and Technology, Sichuan University, Chengdu 610064, China
The purpose of this study was to evaluate the effect of atmo-
sphere on the phase transformation in hydroxyapatite (HA)
coatings during heat treatment by varying the atmosphere
in the furnace pipe. Heat treatment always increased the
crystallinity of HA coatings regardless of the kind of atmo-
sphere. Water molecules in atmosphere further promoted
HA recrystallization during heat treatment. In a dry atmo-
INTRODUCTION
Plasma-sprayed hydroxyapatite (HA) coatings have
been applied to dental implants and hip joint prosthe-
ses extensively because of their ability to form direct
bonding with surrounding bone and early fixation of
implants. The plasma-spraying procedure unavoid-
ably induces partial amorphorization and decomposi-
tion of starting HA powder as a result of the high
temperature of the plasma flame.1,2 Therefore, besides
crystalline HA there is some amorphous apatite and
other calcium phosphates [mainly tricalcium phos-
phate (TCP) and tetracalcium phosphate (TTCP)] in
as-received HA coatings. The crystalline HA usually
biodegrades much slower than amorphous apatite,
TCP, and TTCP.3 Thus, it is desirable to increase the
crystallinity of HA coating and eliminate the fast-
degraded calcium phosphates. There are two ways to
increase the crystallinity of HA coatings: varying
plasma-spraying conditions and post-heat treatment.
However, in vitro and in vivo studies have demon-
strated that varying spraying conditions increase the
crystallinity of HA coating, but do not increase the
stability of HA coatings in a biologic environment.4,5
A suitable method to enhance both crystallinity and
stability of HA coatings in physiologic environments
is post-heat treatment.6 Besides increasing the crystal-
linity of HA coatings, post-heat treatment also induces
the transformation of other calcium phosphates into
HA in a suitable atmosphere. The results of heat treat-
*To whom correspondence should be addressed.
Journal of Biomedical Materials Research, Vol. 34, 15–20 (1997)
q 1997 John Wiley & Sons, Inc.
15
sphere, tricalcium phosphate (TCP) and tetracalcium phos-
phate (TTCP) were more stable than HA, so heat treat-
ment could not convert them into HA. However, in a hu-
mid atmosphere, heat treatment would transform TCP and
TTCP into HA by hydrolytic reactions. q 1997 John Wiley &
Sons, Inc.
ment depend on several factors such as temperature,
time, and atmosphere of treatment. Thoroughly under-
standing these factors is very important for obtaining
HA coatings with long-term stability.
Temperature dependence of recrystallization of the
amorphous phase in HA coatings during heat treat-
ment has been discussed in an earlier paper.11 In this
study we investigated the effect of the atmosphere
environment on the phase transformations in HA coat-
ings during heat treatment at 6308C.
MATERIALS AND METHODS
The starting HA powder used for coatings was well-
crystallized pure HA, which was synthesized in
our laboratory. HA coatings were made on the sand-
blasted surfaces of the titanium plates by plasma spray-
ing (METCO MN Plasma System and AR2000 Robot).
The thickness of HA coatings was about 50 em. The
HA-coating specimens were divided into five sets and
heated separately in vacuum (21025 mm Hg), oxygen,
nitrogen, humid oxygen, and humid nitrogen at 6308C
for 1 h. The heat treatment equipment capable of vary-
ing the atmosphere is shown in Figure 1. The purity
of both bottled oxygen and bottled nitrogen was about
99.7%. Humid oxygen (or nitrogen) was supplied by
oxygen (or nitrogen) passing through a 958C water
bath. The heat treatment was finished in a pipe furnace.
Specimens were put in a quartz furnace pipe, which
was then evacuated to p1025 mm Hg during vacuum
CCC 0021-9304/97/010015-06
 3739N1 / c104-16 / 01-24-97 13:20:00
16
Figure 1. Schematic drawing of the heat treatment setup.
heat treatment. The corresponding atmosphere passed
through the quartz pipe in four other kinds of heat
treatments.
All of the specimens were examined by X-ray diffrac-
tion (XRD) before and after heat treatment. Phase
transformation analysis was performed by comparing
the XRD patterns after heat treatment with those before
heat treatment. Peak heights of 29.88, 2u, and 30.78,
2u, were measured and used as an indication of the
fractions of TTCP and TCP in the HA coating. Varia-
tions in peak heights before and after heat treatment
indicated phase transitions induced by the heat
treatment.
RESULTS
As-received HA coatings
The starting HA powder for coating was pure and
well crystallized. Its XRD pattern was composed of
only sharp peaks of crystalline HA [Fig. 2(a)]. How-
ever, the XRD pattern of as-received HA coating [Fig.
2(b)] was composed of sharp peaks of crystalline
phases and the diffusion maximum of an amorphous
phase. These crystalline phases are HA, a-TCP, and
TTCP. Besides the appearance of additional peaks of
a-TCP and TTCP, HA peaks for the as-received coating
also broadened and decreased by comparison with
those for the starting HA powder.
Crystallinity of HA coatings
The crystallinity of an HA coating can be estimated
from a comparison of the height of the strongest line
CHEN ET AL.
(2u 5 31.748) of the HA-coating XRD pattern with that
of the standard HA powder.1,6 The variation in peak
height after heat treatment reflected the change in crys-
tallinity of the HA coating. The crystallinities of HA
coatings before and after heat treatment are shown in
Figure 3. Heat treatment always led to crystallinity
increment regardless of the kind of atmosphere. The
crystallinity of HA coating increased by nearly a factor
of 3 after heating in vacuum. More increase in crystal-
linity was observed after heat treatment in oxygen or
introgen. There was no obvious difference in crystallin-
ity between the HA coatings heated in oxygen and
nitrogen. The HA coatings treated in humid oxygen
(or nitrogen) showed the highest crystallinity. Nearly
all the amorphous calcium phosphate had converted
to crystalline HA.
TCP and TTCP
Figure 4 shows the effect of atmosphere on the con-
version of additional phases during heat treatment.
Before heating there was a certain amount of TCP and
TTCP in HA coating. The heat treatment in vacuum
did not have much effect on the additional phases in
HA coating. TTCP and TCP in the HA coating reduced
only a little after heating in vacuum. After heat treat-
ment in oxygen or nitrogen, most of the additional
calcium phosphates were converted to HA. If the heat
treatment was finished in a humid atmosphere, the
amount of TCP further reduced and only trace TTCP
was resolved. The transformation rate of TTCP to HA
in the existence of moisture was always faster than
that of TCP.
 3739N1 / c104-17 / 01-24-97 13:20:00

EFFECT OF ATMOSPHERE ON PHASE TRANSFORMATION 17

Figure 3. Crystallinity of HA coatings before and after heat

treatment. NT 5 not heated; V 5 vacuum; N 5 nitrogen;
O 5 oxygen; HN 5 humid nitrogen; HO 5 humid oxygen.

DISCUSSION

Amorphous and additional phases

The starting HA powder for coatings was pure and
well-crystallized HA. No other calcium phosphate
phases could be resolved in it. The amorphous phase
and the additional calcium phosphate phases in as-
received HA coatings could only result from the
plasma-spraying process.
In fact, plasma spraying is a high-temperature pro-
cess. The temperature of the plasma flame can be as
high as over 10,0008C. In dry, high-temperature condi-
tions, as was usually the case established during
plasma spraying, HA is unstable and will decompose
into TCP, TTCP, and CaO.3 Therefore, in most as-re-
ceived HA coatings there are always additional phases.
Possible processes of decomposition in dry, high-tem-
perature conditions can be expressed as the follow-
ing reactions:
2Ca5(PO4)3(OH) R 2Ca3(PO4)2 1 Ca4P2O9 1 H2O (1)
and/or

Figure 2. XRD patterns: (a) HA powder; (b) as-received

HA coating.
Figure 4. Effect of atmosphere on the conversion of ad-
ditional phases to HA. NT 5 not heated; V 5 vacuum;
N 5 nitrogen; O 5 oxygen; HN 5 humid nitrogen;
HO 5 humid oxygen.
 3739N1 / c104-18 / 01-24-97 13:20:00

18 CHEN ET AL.

2Ca5(PO4)3(OH) R 3Ca3(PO4)2 1 CaO 1 H2O (2) recrystallization of an amorphous phase because of the
exponential dependence of D on Df and T. At room
The amount of additional phases in as-sprayed HA
temperature the phase transformation of the amor-
coatings depends on the plasma-spraying parameters
phous apatite in the HA coatings is too slow to observe.
and atmosphere. The moisture in the atmosphere can
During heat treatment the atoms gained high kinetic
decrease decomposition of HA, because HA is more
energy and diffused much faster than in room tempera-
stable than TCP and TTCP in a humid environment.3,13
ture. The faster diffusion of the atoms in heat treatment
For example, HA began to decompose at 8008C in vac-
speeded up the phase transition from amorphous apa-
uum while the HA structure is thermally stable up to
tite to crystalline HA. This is why heat treatment can
12508C in air because of the moisture in air.7
increase the crystallinity of an HA coating regardless
The amorphous phase indicated by the diffusion
of the kind of atmosphere. The higher the temperature
maximum in Figure 2(b) resulted from the melting of
of heat treatment, the higher the crystallinity of the
HA particles in plasma flame and the following fast
HA coatings, as observed in a preceding investigation
cooling. The HA particles injected in high-temperature
by these authors.11
plasma flame usually were partly molten. The molten
The amount of moisture in atmosphere increased
HA was quenched when HA particles impacted with
in the order of vacuum, oxygen (or nitrogen), humid
the cold titanium substrate (usually below 1508C). Be-
oxygen (or nitrogen). Figure 3 shows that the crystallin-
cause the cooling rate could be as high as 1088C/sec,10
ities of the HA coatings after heat treatment also in-
the molten HA did not have enough time to recrystal-
crease in the same order. The specimens treated in
lize. As a result, most of the molten HA solidified into
vacuum had the lowest crystallinity. The highest crys-
an amorphous phase, and only a little of it recrystal-
tallinity was obtained after heating in humid atmo-
lized into crystalline HA.
sphere.
The crystalline HA in the as-sprayed HA coatings
Exchanging oxygen for nitrogen made little differ-
included two parts: unmolten inner cores of HA parti-
ence in the results of heat treatment (Fig. 3). This indi-
cles and recrystallized HA. Certainly, the unmolten
cated that it was the water molecules in the atmosphere
inner cores of the HA particles remained unchanged
rather than the intrinsic characteristic of oxygen or
in their crystalline state. Their crystalline sizes should
nitrogen to promote recrystallization of the amorphous
be about the same as that in the starting HA powder.
apatite in HA coatings.
Crystallites of recrystallized HA in as-sprayed coatings
Obviously, water molecules played an important role
were superfine grains because the crystallites did not
in promoting the transformation of amorphous apatite
have enough time to grow up during the fast solidifi-
to crystalline HA. Water molecules easily reacted with
cation of the molten HA particles. The superfine recrys-
amorphous apatite and OH groups filled the lattice va-
tallized crystallites in HA coatings reduce the average
cancies which resulted from the dehydration of HA dur-
size of the HA crystallites. As a result, the average size
ing plasma spraying. Removing the lattice vacancies is
of the HA crystallites in as-sprayed HA coatings is
beneficial to the movement of apatite molecules. This
much less than that of the starting HA powder. Accord-
means that the activation energy Df for the molecule
ing to the well-known Scherrer formula,8 decrease of
diffusion decreased after the missing OH sites were
the crystalline size should lead to the broadening of
again occupied by OH groups. The decrease in the acti-
HA peaks in XRD patterns. By comparison with Figure
vation energy led to an increase in diffusion coefficient
2(a), the HA peaks of the HA coating in Figure 2(b)
D and promoted the recrystallization of the amorphous
obviously broadened.
apatite. Another effect of water molecules on transfor-
mation was a reduction in free energy of HA. This effect
Transformation of amorphous to crystalline HA also promotes transformation. Thus, at the same tem-
perature, heat treatment in a humid atmosphere can in-
An amorphous phase is metastable and always tends crease the crystallinity of the HA coating more signifi-
toward recrystallization.9 The recrystallization of an cantly than in a dry atmosphere.
amorphous pahse is usually a diffusion-controlled pro-
cess. The nucleation rate and growth velocity of the
crystallites are both in proportion to the diffusion coef- Conversion of TCP and TTCP to HA
ficient of atoms. The diffusion coefficient can be repre-
sented as: In dry, ambient temperatures, TCP and TTCP are
more stable than HA.3 Therefore, heat treatment in
D y exp(2Df/kT )
vacuum, in which there was almost no moisture, could
where T, Df, and k represent the absolute temperature, not transform TCP and TTCP into HA, as shown in
activation energy for the atom diffusion, and Boltz- Figure 4. On the contrary, some HA was decomposed
man’s constant, respectively. The temperature and acti- into TCP and TTCP by vacuum heat treatment.14 If the
vation energy are the most determining factors for ambient atmosphere contains some water vapor, the
 3739N1 / c104-19 / 01-24-97 13:20:00

EFFECT OF ATMOSPHERE ON PHASE TRANSFORMATION 19

situation will be quite different. The hydrolysis reac-

tions, which are the cement-setting reactions in self-
setting calcium phosphate cements, will lead to the
formation of HA. As shown by the solubility phase
diagram for ternary system Ca(OH)2-H3PO4-H2O in the
range of pH 5–10, HA is the most stable.13 TCP and
TTCP can be converted into more stable HA by the
following hydrolytic reactions:

3Ca4P2O91 3H2O R 2Ca5(PO4)3(OH) 1 2Ca(OH)2 (3)

and

5Ca3(PO4)2 1 3H2O R 3Ca5(PO4)3(OH) 1 H3PO4 (4)

After heat treatment in bottled oxygen (or nitrogen),
most of TTCP and TCP were converted to HA. This
result demonstrated that hydrolytic reactions (3) and
(4) in HA coatings could be carried out even in atmo-
spheres with only small amounts of moisture. Reaction
(3) is faster than reaction (4) for two reasons13: TTCP
is less stable than TCP because the isotherm of TCP
lies below that of TTCP, and the structural similarity
of TTCP to HA promotes the rate of HA crystal growth.
Thus, in an HA coating in which both TCP and TTCP
exist, the rate of dissolution of TTCP is usually greater
than that of TCP. This conclusion is in good agreement
with the results of heat treatment shown in Figure 4.
Before heating, the peak height of TTCP was nearly
equal to that of TCP. But after heat treatment in a
humid atmosphere only trace TTCP could be resolved
while significant amounts of TCP still remained in the
HA coating.
Hydrolysis reactions (3) and (4) together were oppo-
site to decomposition reaction process (1). When an
overly high plasma-spraying power was used in the
coating process, some CaO was formed in the HA
coating by reaction (2). To identify whether the CaO
also could be converted into HA by the opposite de-
composition reaction process (2) with the existence of
small amounts of moisture, the b-TCP powder and
CaO powder were thoroughly mixed in a molar ratio
of 3:1 and heated at 6508C for 1 h in vacuum and in
air, separately. The XRD patterns of the mixture after
heating in vacuum and in air are shown in Figure
5(a,b). There are no HA peaks in the XRD pattern of
Figure 5(a). This means that the vacuum heat treatment
could not induce the formation of HA for lack of water
molecules. Heat treatment in air converted almost all
the TCP into HA, but most of the CaO in the mixture
remained unchanged. This experimental result showed
that the existence of water molecules seldom promoted
the opposite reaction process (2); rather, it first induced Figure 5. XRD patterns of the mixture of TCP and CaO:
hydrolytic reaction (4). CaO converted into HA mostly (a) vacuum heat treatment, (b) air heat treatment.
by the following processes:

CaO 1 H2O R Ca(OH)2 (5)

followed by
 3739N1 / c104-20 / 01-24-97 13:20:01
20
5Ca(OH)2 1 3H3PO4 R Ca5(PO4)3OH 1 9H2O (6)
The H3PO4 in reaction (6) was the product of reaction
(4). The reaction rate of process (6) was much slower
than that of process (4). As shown in Figure 5(b), reac-
tion (4) had been finished in the heat treatment period
of the mixture, while process (6) was still in its begin-
ning. Perhaps the CaO in the mixture could be com-
pletely converted into HA if the heat treatment time
were long enough.
CONCLUSIONS
Phase transformation during heat treatment was de-
termined by the relative stability of each kind of cal-
cium phosphate in the HA coatings. The relative stabil-
ity of these calcium phosphates varied with the
atmosphere. The effect of the atmosphere on the phase
transformation in HA coatings during heat treatment
was studied by varying the atmosphere in the furnace
pipe. The results were as follows:
1. Amorphous calcium phosphate is a metastable
phase and always tends toward recrystallization.
Heat treatment accelerated the recrystallization
of the amorphous calcium phosphate and always
led to the crystallinity increment of HA coatings
regardless of the kind of atmosphere.
2. Water molecules in atmosphere during heat treat-
ment promoted HA crystal growth by decreasing
both the activation energy Df for the molecule
diffusion and the free energy of the HA.
3. Heat treatment in a dry atmosphere could not
convert TCP of TTCP into HA, because the former
is more stable than HA in an atmosphere with-
out moisture.
4. Heat treatment in a humid atmosphere could
transform TTCP and TCP into HA by hydrolytic
reactions (3) and (4).
References
1. J. Chen, J. G. C. Wolke, and K. de Groot, ‘‘Microstruc-
ture and crystallinity in hydroxyapatite coatings,’’ Bio-
materials, 15, 396–399 (1994).
CHEN ET AL.
2. S. R. Radin and P. Ducheyne, ‘‘Plasma spraying induced
changes of calcium phosphate ceramic characteristics
and the effect on in vitro stability,’’ J. Mater. Sci. Mater.
Med., 3, 33–42 (1992).
3. K. de Groot, C. P. A. T. Klein, J. G. C. Wolke, and
J. M. A. de Blieck-Hogervorst, ‘‘Chemistry of calcium
phosphate bioceramics,’’ in CRC Handbook of Bioactive
Ceramics, Vol. II, Takao Yamamuro, L. Hench, and J.
Wilson (eds.), CRC Press, Boca Raton, Florida, 1990,
pp. 3–16.
4. S. H. Maxian, J. P. Zawadsky, and M. G. Dunn, ‘‘Evalua-
tion of amorphous vs. crystalline hydroxyapatite coat-
ings on smooth and rough titanium: in vitro and in vivo
studies,’’ in Transactions of the Fourth World Biomaterials
Congress, Berlin, 1992, pp. 101.
5. P. Frayssinet, F. Tourenne, N. Rouquet, I. Primout, and
J. S. Hanker, ‘‘Effect of crystallinity of HA-coatings on
cell proliferation rate,’’ in Transactions of the Fourth World
Biomaterials Congress, Berlin, 1992, p. 304.
6. C. P. A. T. Klein, J. G. C. Wolke, J. M. A. de Blieck-
Hogervorst, and K. de Groot, ‘‘Calcium phosphate
plasma-sprayed coatings and their stability: An in vivo
study,’’ J. Biomed. Mater. Res., 28, 909–917 (1994).
7. J. Weng, X. Liu, X. Zhang, and X. Ji, ‘‘Thermal decompo-
sition of hydroxyapatite structure induced by titanium
and its dioxide,’’ J. Mater. Sci. Lett., 13, 159–161 (1994).
8. H. P. Klug and L. E. Alexander, X-Ray Diffraction Proce-
dures for Polycrystalline and Amorphous Material, 2nd Ed.,
Wiley, New York, 1971.
9. R. Zallen, The Physics of Amorphous Solids, Wiley, New
York, 1983.
10. C. Moreau, P. Cielo, M. Lamontagne, S. Dallaire, J. Kra-
pez, and M. Vardelle, ‘‘Temperature evolution of
plasma-sprayed niobium particles impacting on a sub-
strate,’’ in Thermal Spray Research and Applications, T. F.
Bernecki (ed.), ASM International, Materials Park, Ohio,
1990, pp. 19–26.
11. J. Chen, Y. CaO, J. Feng, and X. Zhang, ‘‘Crystallization
of amorphous phase in HA coatings during heat treat-
ment,’’ in Biomedical Materials Research in the Far East
(I), Xingdong Zhang and Yoshito Ikada (eds.), Kobunshi
Kankokai, Kyoto, Japan, 1993, pp. 17–18.
12. J. Weng, Y. CaO, J. Chen, and X. Zhang, ‘‘Significance
of water promoting amorphous to crystalline conver-
sion of apatite in plasma sprayed coatings,’’ J. Mater.
Sci. Lett., 14, 211–213 (1995).
13. L. C. Chow, ‘‘Development of self-setting calcium phos-
phate cements,’’ J. Ceram. Soc. Jpn., 99, 954–964 (1991).
14. F. Brossa, A. Cigada, R. Chiesa, L. Paracchini, and C.
Consonni, ‘‘Post-deposition treatment effects on hy-
droxyapatite vacuum plasma spray coatings,’’ J. Mater.
Sci. Mater. Med., 5, 855–857 (1994).
Received September 28, 1995
Accepted February 16, 1996