Biomaterials 19 (1998) 2291 2296

Critical ageing of hydroxyapatite solgel solutions

Cameron S. Chai*, Karlis A. Gross, Besim Ben-Nissan
Department of Chemistry, Materials and Forensic Sciences, University of Technology, Sydney Box 123, Broadway, NSW 2007, Australia
Received 12 February 1998; accepted 25 May 1998

Abstract
It has been established that hydroxyapatite coatings can be produced using an alkoxide based solgel technique. Previous work
showed that in addition to hydroxyapatite other phases including CaO were observed. A critical factor in determining the
composition of the final coating is the period between solution preparation and coating deposition. The current study addresses how
the ageing time affects the composition of the coating and the time required before a solution can be used to deposit monophasic
coatings.
Powders and coatings were produced with ageing times up to one week and examined with X-ray diffraction, thermal gravimetric
analysis and differential thermal analysis. By depositing coatings after various ageing times, it was found that a solution ageing time of
at least 1 day was required before monophasic hydroxyapatite coatings could be deposited. Thermogravimetric analysis showed that
heating of the gel in air or in nitrogen both produced hydroxyapatite. Thermogravimetric analysis could be used as a simple and
effective tool for identifying the critical ageing time required to produce high-purity hydroxyapatite coatings. ( 1998 Published by
Elsevier Science Ltd. All rights reserved
Keywords: Hydroxyapatite; Ageing; Solgel; Coating

1. Introduction
The rationale in using hydroxyapatite coatings as
a means of fixation for orthopaedic and dental implants
has been known since the early 1980s [1] Several
methods such as dip coating [2], sputter deposition [3],
pulsed laser deposition [4], electrophoretic deposition
[5, 6] and thermal spray techniques [7, 8] have been
documented over the years. However, of these techniques
only plasma spraying has enjoyed commercial success.
These coatings can be applied to a thickness of between
50 and 200 lm.
An alternative method for hydroxyapatite coating deposition is solgel. This produces coatings comparable in
thickness with sputter deposition which are typically less
than 1 lm. The term solgel encompasses any process of
producing ceramic materials (single and mixed oxides, as
well as non-oxides, e.g. nitrides) from solutions. The
solgel process was first identified by Ebelman [9] and
has since been used to produce ceramic powders [10],

* Corresponding author.

coatings [11, 12] and bulk materials including glasses

[13].
Solgel offers a number of advantages over other coating techniques such as; increased homogeneity as mixing
occurs on the atomic level; reduced sintering temperatures due to small particle size; ability coat complex
shapes easily; and numerous deposition techniques that
can be employed to produce coatings [14]. It was originally used to produce simple oxides and glasses, however,
more recently its application has included the production
of mixed oxides and nitrides.
Solgel synthesis involves preparation of a solution
from alkoxides, metal salts or other suitable precursor,
such as nitrates. To produce a coating, the solution can
then be deposited by either dip, spin or spray coating.
The solution or coated substrate is then exposed to water
for hydrolysis. During this process hydroxides or hydrated oxides form and gelation occurs to form a threedimensional network. Further heating of the resultant gel
removes residual organic material and induces conversion to the oxide state under oxidising conditions [15].
The development of hydroxyapatite using the solgel
route is based on the initial work of Masuda et al. on

0142-9612/98/$See front matter ( 1998 Published by Elsevier Science Ltd. All rights reserved.
PII S 0 1 4 2 - 9 6 1 2 ( 9 8 ) 0 0 1 3 8 - 0

2292

C.S. Chai et al. / Biomaterials 19 (1998) 22912296

powder production [16], which was applied with further

modification for coating production using the spin-coating technique [17]. In producing hydroxyapatite coatings X-ray diffraction revealed a contaminant phase after
immediate use of the solgel. This was later identified as
calcium oxide [18]. Further work indicated that to produce high-purity hydroxyapatite coatings, solgel solutions must be aged for a period of time before deposition
onto substrate materials and subsequent hydrolysis and
heat treatment [19]. The present study is aimed at determining the ageing time required for monophasic hydroxyapatite coating production.

2. Experimental procedure
The current work employs a conventional alkoxide
processing methodology similar to that used by Masuda
et al. [16] who successfully synthesised hydroxyapatite
powders via the solgel technique.
A calcium alkoxide solution was prepared by dispersing 1.5]10~3 mol of calcium diethoxide (Kojundo
Chemical Lab., Saitama, Japan) in ethanol and then
dissolving the calcium diethoxide with ethanediol (BDH,
England) while stirring vigorously. A phosphorus alkoxide solution was then prepared by diluting triethyl phosphite (Sigma Chemical Co., St Louis, USA) in ethanol
(BDH, Kilsyth, Australia). When the calcium diethoxide
was completely dissolved, a stoichiometric quantity of
the phosphorus alkoxide solution was added dropwise to
the calcium alkoxide solution. Vigorous agitation was
maintained throughout the addition process and for
a further 10 min thereafter. Due to the hygroscopic nature of the reactants, the preparation was conducted in
a glove box containing a dry nitrogen atmosphere and
the vessels were closed to avoid any volatilization. The
resultant solutions were then aged for increasing periods
for the production of gels and coatings.
After ageing, gels for thermal analysis were prepared
by pouring the appropriate solution into a petri dish and
hydrolysing at 70C in a convection oven (Labec, Sydney, Australia). The resultant gels were then crushed
using an agate mortar and pestle and analysed using
thermogravimetric analysis (TGA) and differential thermal analysis (DTA) using a SDT 2960 simultaneous
thermal analyser (TA Instruments, Newcastle, USA). The
heating regime involved heating at 10C min~1 to 500C,
followed with a 15 min hold and then heating at
3.33C min~1 to 1200C. The heating profile was used to
simulate the heat treatment conditions for coating production. A stagnant air atmosphere was employed for
most of the experiments. Flowing air and nitrogen atmospheres were also used to determine if there was any effect
of firing atmosphere.
Single-crystal magnesia (Zirmat, N Bellerica, USA)
was selected as the substrate. This enabled the changes

with temperature to be observed without interference of

the substrate, which normally would be a metallic material. Prior to coating, the substrates were ultrasonically
cleaned twice in acetone and ethanol and dried at 110C
in an air oven. Solutions destined for coatings were
removed from the glove box in septum sealed glass vials.
Deposition involved applying 0.05 ml of solution to the
surface of the substrate, such that the entire upper surface
was covered. Excess solution was removed using a Headway Research (Garland, USA) spin coater (2500 rpm,
10 s spin time). Coated substrates were hydrolysed for
10 min at 70C in an air oven and then prefired to 500C
in a muffle furnace (Ceramic Engineering, Sydney, Australia) for 15 min. It was found that to assist the build
up of coating thickness, pre-firing was required prior
to the deposition of the subsequent layers. This was
necessary as the solvents contained in the solution were
found to disrupt the formation of uniform layers. Hence,
a pre-firing schedule was used as a means of increasing
the stability of previously deposited layers. The coating
process was repeated 5 times to build up the thickness of
the coating after which they were fired to 1000C for
15 min.
Coated samples were analysed by X-ray diffraction
(XRD) using a Siemens D5000 diffractometer (Karlsruhe,
Germany), with a grazing incidence geometry. Scanning
parameters included a scan range of 2940 2#, step size
of 0.02, step time of 5 s and X-ray incident angle of 1.
This configuration was found to be necessary, because
conventional X-ray diffraction penetrated too far into the
sample. In this instance, only the substrate was detected.
By using the grazing incidence XRD, the signal from
coating relative to the substrate is maximised.

3. Results and discussion

In the early stages of this research, the production of
mono-phasic hydroxyapatite coatings with solutions not
aged prior to deposition proved very difficult. However,
when solutions were aged for 7 days prior to coating, the
amount of hydroxyapatite in the coatings was found to
increase. This particular variable then had significant
importance, as it determined the final composition of the
coating. The shorter maturation time for the solution
would increase the efficiency of the process and render it
favourable for an industrial setting.
Thermal gravimetric analysis was used as one of the
characterisation tools. It showed a slow and steady
weight loss through to the hold in temperature at 500C,
(Fig. 1a). The vertical translation observed at 500C can
be attributed to continual weight loss during the
15 min at that temperature. After 500C, samples aged
for less than 4 h maintain constant weight until approximately 680C. As ageing time increases up to 24 h, the
rate of weight loss up to 680C increases gradually. After

C.S. Chai et al. / Biomaterials 19 (1998) 22912296

2293

Fig. 1. Thermal analysis showing (a) TGA and (b) DTA curves for
samples aged for up to 24 h.

Fig. 2. Thermal analysis showing (a) TGA and (b) DTA curves for
samples aged from 1 to 7 days.

680C samples aged less than 24 h all exhibited a more

rapid and larger weight loss compared to samples aged
for longer periods. X-ray diffraction and thermal
gravimetric studies have indicated that this weight loss
for insufficiently aged samples is attributed to unreacted
calcium diethoxide which transforms to calcium carbonate during heating and decomposes to calcium oxide
above 600C. The TGA for samples aged longer than
24 h do not differ significantly from one another (Fig. 2a).
In particular, the total weight loss after 800C remains
fairly constant at 2830 wt% compared to the 50 wt%
weight loss for non-aged solutions. Higher weight losses
such as those observed with solutions aged for less than
24 h would have deleterious effects on coating quality,
since the decomposition of calcium carbonate could increase the susceptibility for pore formation and cracking
through the release of gases and produce non-stoichiometric hydroxyapatite.
TGA curves for samples heated in flowing air and
flowing nitrogen for a sample aged for 1 day, did not
produce any difference (Fig. 3). This suggests that coatings could be manufactured in an oxygen-free environment which is necessary to prevent oxidation of metallic
substrates when they are chosen for use.

Differential thermal analysis of the aged samples begin

with a large endotherm at 100110C which can be
attributed to the evolution of excess alcohol and/or
water. The non-aged sample displays a large exotherm at
approximately 200C, followed closely by two poorly
defined exothermic reactions at 250 and 300C (Fig. 1b).
The final exothermic reaction observed at 400C seems to
be present at all times. For samples aged less than 24 h,
an endothermic reaction at 700C is associated with
a rapid weight loss. Exothermic reactions observed between 200 and 400C are believed to be due to oligomerization and polycondensation.
As ageing time increases, four changes can be observedtwo within the low-temperature regime and the
others between 500 and 700C (Fig. 2b). The two poorly
defined exothermic reactions at 250 and 300C, and the
endothermic reaction at 700C gradually disappear, until
they can no longer be detected after 24 h. This is consistent with TGA results and the reduction of the endotherm at 700C is associated with a smaller weight
loss. Another small-peak appears at about 550C. The
small exothermic response represents crystallization
from an amorphous phase which has also been identified
in thermally sprayed [20, 21], sputtered [22] and pulsed

2294

C.S. Chai et al. / Biomaterials 19 (1998) 22912296

Fig. 3. Thermal analysis showing (a) TGA and (b) DTA curved for
samples heated in stagnant air, flowing nitrogen and air. Flow rates for
nitrogen and air are 140 cc min~1.

Fig. 4. X-ray diffraction for samples aged for (a) up to 24 h and (b) from
1 to 7 days. Peaks labelled with an asterisk represent Laue reflections
from the single-crystal substrate.

laser deposited coatings [23]. The similarity of the crystallization temperature to that reported in thermally
sprayed amorphous calcium phosphate coatings [20]
suggests that there is sufficient hydroxyl species present
for crystallization to hydroxyapatite.
Hence, thermal analysis indicates that the system in
question reaches a stable configuration after 24 h ageing
time. At this point, the exothermic reactions at 250 and
300C have disappeared and the endothermic weight loss
reaction at 700C is no longer detectable.
Thermal analysis showed that samples heated in air or
nitrogen do not produce any notable difference in the
conditions used, suggesting that a non-oxidising atmosphere may be utilized for reactive metals (Fig. 3b).
X-ray diffraction of the samples aged less than 24 h
indicates the presence of hydroxyapatite (JCPDS 9-432)
and calcium oxide (JCPDS 4-777) (Fig. 4a). The superposition of the calcium oxide and hydroxyapatite peak at
32.1 produces a broadening and higher intensity of the
(1 1 2) peak. With aging times less than 8 h, unidentified
peaks appear at 31.2. This does not correspond to tricalcium phosphate and the high-processing temperature
removes the possibility of anhydrous orthophosphoric

acid. It is possible that this peak could arise from a phosphate-rich phase. The amount of hydroxyapatite relative
to calcium oxide is observed to increase as ageing times
approach 24 h, at which point hydroxyapatite is the only
identifiable phase. For ageing times up to 1 week, sols
produce monophasic hydroxyapatite when deposited
(Fig. 4b). Peaks labelled with an asterisk represent Laue
reflections from the single crystal substrate [24].
The key findings in this section of research were the
differing TGA behaviour and XRD determined compositions as a function of ageing time. The total weight loss of
the hydrolysed gel increased with decreasing ageing time.
Also, CaO is found in the coatings with ageing time less
than 24 h. In other work where different reactants were
used CaCO and Ca(OH) was found [25]. The presence
3
2
of these calcium compounds and the loss in weight at
higher temperatures are both indicators suggesting the
lack of a monophasic hydroxyapatite phase.
It was found that a key factor in influencing the final
composition of a coating deposited by solgel techniques
in the current system was the length of ageing time prior
to use for coating, provided firing temperatures remained
below the thermal stability limit. The ageing or maturing

C.S. Chai et al. / Biomaterials 19 (1998) 22912296

period is thus required to allow chemical reactions to

reach completion at room temperature in the solution.
The exact nature of these reactions is under investigation,
however, some insight can be gained from the literature
on calcium phosphate synthesis.
Synthesis of hydroxyapatite through aqeous solution
chemistry has shown that an ageing time is required to
allow complete homogenisation [26] and formation of
a stoichiometric composition [27]. Availability of the
calcium species is one of the rate limiting steps [28] and it
has been reported that it is possible to decrease the
maturation time by using dilute solutions [29]. The application of a lower-concentration solution for solgel
coating, however, results in patchy uneven coatings. This
requires more layers to build up the required thickness.
Thus, it can be concluded from this section of work
that a maturing period is required for the solution before
it can be used to produce hydroxyapatite coatings. Preliminary results suggest that the process is also dependent on the quantity of calcium diethoxide and triethyl
phosphite present. Ageing time increases with increasing
concentrations of these reactants. It is possible that the
maturing time could be reduced by reducing the amount
of solvents used in the solution preparation, but this may
also adversely effect the overall solution chemistry and
the resultant quality of the coatings.

[2]
[3]

[4]

[5]

[6]
[7]

[8]

[9]
[10]
[11]
[12]

[13]

4. Conclusions
It was found that coatings manufactured with an alkoxide based solgel process without ageing consists of
calcium oxide in addition to hydroxyapatite. A period
greater than 24 h is required for the production of monophasic hydroxyapatite. The solution must be aged until
there is no more abrupt weight loss between 680 and
750C upon heating. Thermal analysis showed that samples heated in air or nitrogen do not produce any notable
difference in the conditions used, suggesting that a nonoxidising atmosphere may be utilized for reactive metals.
Thermal gravimetric analysis in conjunction with X-ray
diffraction can be used as an effective method to identify
whether the solution has been sufficiently aged.

[14]
[15]
[16]

[17]

[18]

[19]

[20]

Acknowledgements
The authors are grateful for discussions on grazing
angle with Dr. R. Cheary and the use of X-ray facilities in
the Microstructural Analysis Unit.

References

[21]

[22]

[23]
[24]

[1] Ducheyne P, Hench LL, Kagan A, Martens M, Burssens A,

Mulier JC. Effect of hydroxyapatite impregnation on skeletal

2295

bonding of porous coated implants. J Biomed Mater Res

1980;14:22537.
Lacefield WR. Hydroxyapatite coatings. Ann NY Acad Sci
1988;523:7280.
Ong JL, Lucas LC, Lacefield WR, Rigney ED. Structure, solubility and bond strength of thin calcium phosphate coatings produced by ion beam sputter deposition. Biomaterials 1992;
13(4):24954.
Cotell CM, Chrisey DB, Grabowski KS, Sprague JA, Rossett CR.
Pulsed laser deposition of hydroxylapatite thin films on Ti
6Al4V. J Appl Biomater 1992;3:8793.
Ducheyne P, van Raemdonck W, Heughebaert JC, Heughebaert
M. Structural analysis of hydroxyapatite coatings on titanium.
Biomaterials 1986;7:97103.
Zhitomirsky I, Gal-Or L. Electrophoretic deposition of hydroxyapatite. J Mater Sci: Mater Med 1997;8(4):2139.
de Groot K, Geesink RGT, Klein CPAT, Serekian P. Plasma
sprayed hydroxyapatite coatings of hydroxylapatite. J Biomed
Mater Res 1987;21:137581.
Berndt CC, Haddad GN, Farmer AJD, Gross KA. Thermal
spraying for bioceramic applicationsa review. Mater Forum
1990;14:16173.
Ebelman M. Untersuchungen uber die Verbindung der Borsaure
und Kieselsaure mit Aether. Ann Chim Phys 1846;57:31955.
Mazdiyasni KS. Powder synthesis from metal-organic precursors.
Ceram Int 1982;8(2):4258.
Guglielmi M. Solgel coatings on metals. J SolGel Sci Technol
1997;8:4439.
Dislich H. Thin films from the solgel process. In: Klein LC,
editor. Solgel technology for thin films, fibers, preforms, electronics and specialty shapes. New Jersey: Noyes Publications,
1988:5070.
Thomas IM. Multicomponent glasses from the solgel process.
In: Klein LC, editor. Solgel technology for thin films, fibers,
preforms, electronics and specialty shapes. New Jersey: Noyes
Publications, 1988:215.
Uhlmann DR, Teowee G, Boulton J. The future of solgel science
and technology. J SolGel Sci Technol 1997;8:108391.
Hench LL, West JK. The solgel process. Chem Rev 1990;90:3372.
Masuda Y, Matubara K, Sakka S. Synthesis of hydroxyapatite
from metal alkoxides through solgel technique. J Ceram Soc
Japan 1990;98:8494.
Chai CS, Ben-Nissan B, Pyke S, Evans L. Solgel derived hydroxpatite coatings for biomedical applications. In: Sudarshan TS,
Ishizaki K, Takata M, Kanata K, editors. Surface Modification
Technologies, vol. 7. 1993:50923.
Ben-Nissan B, Chai CS. Solgel derived bioactive hydroxyapatite
coatings. In: Kossowsky R, Kossovsky N, editors. Advances in
materials science and implant orthopaedic surgery. Netherlands:
Kluwer Academic Publishers, 1995:26573.
Ben-Nissan B, Chai CS, Gross KA. Effect of solution ageing on
solgel hydroxyapatite coatings. In: Sedel L, Rey C, editors.
Bioceramics, vol. 10. Paris: Elsevier Science, 1997:1758.
Gross KA, Gross V, Berndt CC. Thermal analysis of the amorphous phase in hydroxyapatite coatings. J Amer Cer Soc 1998;
81(1):10612.
Weng J, Liu XG, Li XD, Zhang XD. Intrinsic factors of apatite
influencing its amorphization during plasma-spray coating. Biomaterials 1995;16(1):3944.
Ong JL, Lucas LC. Post-deposition heat treatments for ion beam
deposited calcium phosphate coatings. Biomaterials 1994;
15(5):33741.
Cottell CM. Pulsed laser deposition and processing of biocompatible hydroxylapatite thin films. Appl Surf Sci 1992;69:1408.
Eatough MO, Goehner RP. The effects of using long soller slits as
parallel beam optics for GIXRD on diffraction data. In: Gilfrich
GV, Huang TC, Jenkins R, Cev Noyan I, Smith DK, Predecki

2296

C.S. Chai et al. / Biomaterials 19 (1998) 22912296

PK, editors. Advances in X-ray analysis. New York: Plenum

Press, 1994:16773.
[25] Kordas G, Trapalis CC. Fourier transform and multi dimensional
EPR spectroscopy for the characterization of solgel derived
hydroxyapatite. J SolGel Sci Technol 1997;9:1724.
[26] Eanes ED, Gillessen IH, Posner AS. Intermediate states in the
precipitation of hydroxyapatite. Nature 1965;208:3657.

[27] Eanes ED, Meyer JL. The maturation of crystalline calcium

phosphates in aqueous suspensions at physiologic pH. Calcif Tiss
Res 1977;23:25969.
[28] Jarcho M. Calcium phosphate ceramics as hard tissue prosthetics.
Clin Orthop 1981;157:25978.
[29] Boskey AL, Posner AS. The formation of hydroxyapatite at low
supersaturations. J Phys Chem 1976;80:446.