J. Am. Ceram. Soc.
, 91 [5] 1528–1533 (2008)
Journal
Early Stages of Calcium Phosphate Formation on Bioactive Borosilicate
Glass in Aqueous Phosphate Solution
Yadong Li,z,y Mohamed N. Rahaman,*,w,z B. Sonny Bal,z Delbert E. Day,*,z,J and Qiang Fuz
z
Department of Materials Science and Engineering, University of Missouri-Rolla, Rolla, Missouri 65409
y
School of Materials Engineering, Soochow University, Suzhou 215021, China
z
Department of Orthopaedic Surgery, University of Missouri-Columbia, Columbia, Missouri 65212
J
Graduate Center for Materials Research, University of Missouri-Rolla, Rolla, Missouri 65409
The conversion of bioactive glass to a calcium phosphate material,
typically hydroxyapatite (HA), by solution–precipitation reac-
tions in aqueous phosphate solution, has been commonly reported.
This paper describes the structural and compositional charac-
teristics of the calcium phosphate material formed during the
early-stage conversion (o5 h) of a borosilicate glass (designated
H12) in aqueous phosphate solution. Disks of H12 glass were
reacted with 0.25M K2HPO4 solution with a starting pH 5 7.0
at 371C. The structure and composition of the product layer
were characterized using thin film X-ray diffraction,
scanning electron microscopy, and Fourier transform infrared
reflectance spectroscopy. For reaction times o5 h, the results
indicated the formation of brushite (CaHPO4 . 2H2O) with
coarse, plate-like crystals. In comparison, the formation of
HA with small needle-like crystals was found at later times.
This early-stage formation of brushite has not been reported
in previous studies of converting bioactive silicate and borate
glasses in aqueous phosphate solution.
I. Introduction
C ERTAIN glasses, described as bioactive glasses, have been
observed to react chemically with aqueous phosphate solu-
tion, such as the body fluids in animals and humans, forming a
hydroxyapatite (HA) layer that bonds strongly to hard or soft
tissues.1,2 Both silicate-based and borate-based glasses, as well as
glass–ceramics based on the silicate compositions, have been
observed to be bioactive.3–6 There is a growing interest in the use
of these bioactive glasses and glass-ceramics for dental and
medical applications, such as the repair and regeneration of
damaged or disease bone, fixation of long bone fractures, and
hip- and knee-joint revision surgery.4–6 These glasses and glass–
ceramics are osteoconductive as well as osteoinductive, and have
the ability to enhance growth and differentiation of cells, so they
are attractive materials for use as scaffolds in tissue-engineered
regeneration of bone.7,8
The mechanisms by which bioactive silicate and borate glasses
convert to HA in an aqueous phosphate solution have been dis-
cussed in detail elsewhere.4,5,9,10 In the case of silicate-based
bioactive glass, such as the well-researched 45S5 composition, it
is well established that an initial step in the reaction is the for-
mation of a thin SiO2-rich gel layer (Ba few micrometers thick)
on the 45S5 glass surface by ion-exchange reactions. Further
dissolution of ions from the glass and their diffusion through the
D. Johnson—contributing editor
Manuscript No. 22944. Received March 16, 2007; approved August 26, 2007.
*Member, The American Ceramic Society.
w
Author to whom correspondence should be addressed. e-mail: rahaman@umr.edu
1528
DOI: 10.1111/j.1551-2916.2007.02140.x
r 2008 The American Ceramic Society
SiO2-rich gel layer, followed by the reaction between Ca21 ions
from the glass and PO3
4 ions from the surrounding liquid, leads
to the formation of an amorphous calcium phosphate (ACP)
layer on the SiO2 gel layer. The reaction between the glass and
the phosphate solution continues, and at the same time the ACP
converts to HA, giving a structure consisting of an unconverted
glass core, an SiO2-rich layer, an ACP layer, and a growing
outer layer of HA. Eventually, the reaction stops and, with the
completion of the ACP conversion to HA, a composite structure
consisting of an HA layer surrounding a SiO2-rich core is pro-
duced. The conversion of bioactive borate glass to HA follows a
process similar to that for 45S5 glass, but without the formation
of a SiO2-gel layer.9,10 Reaction of the glass with an aqueous
phosphate solution to form ACP directly, coupled with the con-
version of the ACP, leads to a growing HA product layer.
In the present work, the early stages of calcium phosphate
formation on a bioactive borosilicate glass were observed to be
different from those reported for bioactive silicate and borate
glasses. The objective of the work was to characterize the struc-
tural and compositional changes in the early stage of converting
the borosilicate glass to a calcium phosphate material in aque-
ous phosphate solution. Scanning electron microscopy (SEM),
thin-film X-ray diffraction, and Fourier transform infrared
(FTIR) reflectance spectroscopy were used to characterize the
product layer formed on the glass for reaction times up to 1 day.
The results were compared with those of previous work for bio-
active silicate and borate glasses. Bioactive borosilicate glass can
provide alternative materials for biomedical applications such as
bone repair. An advantage is that the dissolution rates and con-
version to a HA-type material can be tailored between those of
the more durable silicate bioactive glasses (such as 45S5) and the
more rapidly reacting borate bioactive glasses. As such, an un-
derstanding of the conversion mechanism for bioactive borosili-
cate glass is also important.
II. Experimental Procedure
The borosilicate glass used in the present work was designated
H12, with a composition (in mol%) of: 40.0 B2O3, 7.5 SiO2, 40.0
CaO, 8.0 Na2O, 2.0Al2O3, and 2.5 P2O5. Developed and used in
previous work,11 the glass was found to be bioactive and to
provide a substrate favorable for cell growth, in addition to
forming a strong bond to titanium alloys used for total joint
replacement.12 The glass was prepared by melting the required
quantities of SiO2 (Fused quartz, Particle Processing & Classi-
fying Corp., Patterson, NJ), CaCO3, Na2CO3, Al2O3, NaH2
PO4  2H2O, and H3BO3 (Reagent grade; Fisher Scientific, St.
Louis, MO) in a platinum/rhodium crucible in air for 2 h
at 13001C. The melt was stirred twice, and then cast between
two steel plates to form sheets. Disks with a rectangular cross
section (approximately 15 mm  10 mm  1 mm), for use in
the glass conversion experiments, were prepared by cutting
May 2008 Early Stages of Calcium Phosphate Formation 1529

(a) 300 (b)

Reaction time = 0.5 h Reaction time = 1 h
1500
250
HA
Brushite
Brushite
200
1000

Intensity
Intensity

150

100 500

50

0
0

20 40 60 80 20 40 60 80
2θ (degree) 2θ (degree)
(c) (d) 2000
2500 Reaction time = 5 h
Reaction time = 3 h
HA
HA
2000 1500 Brushite
Brushite
Intensity

Intensity
1500
1000
1000

500
500

20 40 60 80 10 20 30 40 50 60 70
2θ (degree) 2θ (degree)
(e) 2000
Reaction time = 1 d

1500 HA
Brushite

1000
Intensity

500

10 20 30 40 50 60 70
2θ (degree)
Fig. 1. Thin film X-ray diffraction patterns of H12 glass disks reacted in 0.25M K2HPO4 solution with a starting pH 5 7.0 at 371C for (a) 0.5 h, (b) 1 h,
(c) 3 h, (d) 5 h, and (e) 1 day. The peaks corresponding to those of a reference hydroxyapatite (HA) (JCPDS 72-1243), and a reference calcium hydrogen
phosphate dihydrate (brushite) (JCPDS 72-1240) are indicated.

with a diamond-coated blade, and ground with 600 grit silicon
carbide paper.
The phosphate solution used in the conversion of the glass to
HA was prepared by dissolving K2HPO4 (Reagent grade, Fisher
Scientific) in deionized water to give a solution with a concen-
tration of 0.25M and pH 5 7.0. Glass disks were suspended in
the phosphate solution, and the system was placed in an oven at
371721C. The system was stirred periodically. Approximately
1 g glass and 100 cc of phosphate solution were used, which were
identical to the quantities of reactants used in previous work.9,10
Individual disks were removed from the solution after reaction
times of 0.5, 1, 3, 5, and 24 h, washed three times with distilled
water, then twice with Reagent grade ethanol, and finally dried
overnight at 901C.
The structure and composition of the crystalline product lay-
er on the glass were determined by X-ray diffraction (XRD), in
the thin-film mode (XDS 2000; Scintag, Cupertino, CA), using
CuKa radiation (l 5 0.15406 nm) in a step-scan mode (0.051 2y)
in the range 31–901 2y. Compositional information on the reac-
tion products was also obtained using FTIR reflectance spec-
troscopy (Model 1760-X; Perkin Elmer, Eden Prairie, MN) in
the wavenumber range of 400–1500 cm
1. Each spectrum was
1530
Reflectance (arbitrary units)
P-O str.
P-O str.
P-O str.
1d
5h
1h
1400 1200 1000 800
Wavenumber (cm−1)
Fig. 2. Fourier transform infrared reflectance spectra of H12 glass
disks reacted in aqueous phosphate solution for 1 h, 5 h, and 1 day.
the average of 100 scans at a resolution of 1 cm
1. The reacted
glass disks were coated with gold–palladium and observed in a
SEM (S-570; Hitachi, Tokyo, Japan).
III. Results and Discussion
Figures 1(a)–(e) shows XRD patterns of the H12 glass disks
reacted in the phosphate solution for 0.5, 1, 3, 5, and 24 h,
respectively. Figure 1(a) shows a generally amorphous back-
ground, but with a peak appearing at B11.51 2y. This peak
intensity (taken as the height) increased significantly after a
reaction time of 1 h (Fig. 1(b)), with the appearance of addi-
tional peaks, which continued to increase in intensity after 3 h of
reaction (Fig. 1(c)). These peaks correspond to those of the
mineral brushite, calcium hydrogen phosphate dihydrate, Ca
HPO4  2H2O (JCPDS 72-1240), indicating the formation of cal-
cium hydrogen phosphate during the early-stage conversion
reaction. After 5 h reaction (Fig. 1(d)), the diffraction pattern
became markedly different, containing peaks corresponding to
those of HA, Ca10(PO4)6(OH)2 (JCPDS 72-1243), as well as
peaks of CaHPO4  2H2O. Further reaction for 1 day gave a
product layer with diffraction peaks corresponding to HA
(Fig. 1(e)), and with only a few minor peaks corresponding to
CaHPO4  2H2O. In other studies,13 H12 glass particles or disks,
reacted for longer times (10–20 days) in a similar phosphate so-
lution, gave a diffraction pattern corresponding to HA, with no
detectable peaks corresponding to CaHPO4  2H2O. According
to the present XRD data, the conversion of H12 glass in 0.25M
K2HPO4 solution initially produced a product layer of Ca
HPO4  2H2O, which subsequently converted to HA or was cov-
ered with a growing HA layer.
The formation of HA after reaction times of B5 h was also
revealed by FTIR reflectance spectroscopy. As shown in Fig. 2,
the spectrum for the glass reacted for 5 h started to show the
major resonances14,15 associated with HA, at B600 cm
1 char-
4 , and at B1023 cm

1
acteristic of P–O (bending) in PO3
char-
acteristic of the P–O resonance (stretching) in PO3
4 . These
resonances became more pronounced in the spectra for a reac-
tion time of 1 day.
SEM images of the surfaces of the H12 glass disks reacted
in the phosphate solution for 0.5, 1, and 3 h are shown in
Figs. 3(a)–(f). After 0.5 h reaction, the surface consisted essen-
tially of an amorphous layer with a few coarse, plate-like crys-
tals. The amorphous layer contained numerous cracks typically
produced by capillary-induced drying stresses. As outlined in the
experimental procedure, after reaction, the glass disks were dried
by placing in an oven at 901C, leading presumably to rapid dry-
ing and large capillary stresses. With increasing reaction time
Journal of the American Ceramic Society—Li et al. Vol. 91, No. 5
(1 h), the amount of coarse plate-like crystals increased, so that
after 3 h, the surface was almost completely covered with a lay-
ered structure of plate-like crystals. Higher magnification SEM
images (Figs. 4(a) and (b)) more clearly revealed the layer-
structured, plate-like crystals.
A distinct change in the structure of the reaction product was
P-O bend. observed after 5 h (Figs. 5(a) and (b)), when SEM revealed a
surface of small, needle-like crystals. These needle-like crystals
coarsened with longer reaction time, but the morphology was
essentially the same (Figs. 5(c) and (d)). The needle-like crystals
were similar in structure and morphology to HA formed by
reacting H12 glass for much longer times (10–20 days),13 and to
HA formed by converting bioactive silicate and borate glasses
by a similar process.9,10,16
The present data showed that a crystalline compound, iden-
tified by XRD as calcium hydrogen phosphate dihydrate, Ca
HPO4  2H2O, formed initially on the surfaces of H12 borosili-
600 400 cate glass, before subsequent conversion of the glass to HA in an
aqueous phosphate solution. The conversion of several bioactive
silicate and borate glasses to HA has been observed and char-
acterized, but this is the first time that the formation of a crys-
talline CaHPO4  2H2O compound has been detected in the early
stages of the conversion process. Under conditions similar to
those used in the present experiments, the early-stage conversion
of silicate 45S5 bioactive glass involved the formation of an
SiO2-rich layer, followed by an ACP layer which subsequently
converted to HA.4,5 For bioactive borate glass, the conversion
mechanism resulted in the direct formation of an ACP layer
which subsequently converted to HA.9,10,16
At 371C, it is well established17 that HA is the most stable
calcium phosphate phase in aqueous phosphate solution with
pH 4 4.0–4.5, whereas calcium hydrogen phosphate is the most
stable phase below pH 5 4.0–4.5. In the present experiments
performed at a starting pH 5 7.0, the formation of HA at longer
reaction times (e.g., 4B5 h) is consistent with stability data for
calcium phosphates, as well as with observations for the con-
version of several bioactive silicate and borate glasses. In the
early-stage conversion, however, the difference in the reaction
mechanism of H12 glass, when compared with silicate (45S5)
glass and borate glass, might be related to kinetic factors or to
the influence of the glass composition.
Because phosphoric acid is a triprotic acid, there will be four
species formed in solution,18 depending on the pH: H3PO4,
H2PO
2
3

4 , HPO4 , and PO4 . Of these, HPO4
2

is more stable
above pH 7, whereas PO3
4 is more stable above pH  11.
As HPO2
4 was expected to have the highest concentration in
the starting phosphate solution used in the conversion process
(pH 5 7.0), the formation of CaHPO4  2H2O might be favored
kinetically in the early stage. At present, it is not clear whether
the CaHPO4  2H2O subsequently converted to the more stable
HA or formed an intermediate layer between the unconverted
glass and the growing HA layer. The pH of the solution in-
creased from the starting value of 7.0 to a value of 7.4 during the
first 24 h of the reaction, so this small increase in the pH is not
expected to have a significant influence on the change from
CaHPO4  2H2O to HA.
The concentration of phosphate ions in the solutions used in
the present work (0.25M K2HPO4) was identical to that used in
previous work,9,10 but much higher than that of a simulated
body fluid (SBF).19 As the phosphate ion concentration is
decreased, the rate of formation of a calcium phosphate layer
on the glass becomes slower, but the ability to form a phosphate
layer still remains. On the basis of the slower conversion rate,
the formation of HA might be expected even in the early stages
of converting H12 glass in an SBF, because it is the most stable
calcium phosphate phase at pH values above 4.0–4.5.17 Because
H12 glass contains 2.5 mol% P2O5, the formation of a calcium
phosphate layer might be expected even in deionized water.
Further, experiments are required to characterize the early-stage
formation of the calcium phosphate phase on H12 glass in an
SBF, as well as in deionized water. A calcium phosphate layer
was observed to form on bioactive 45S5 glass in deionized
May 2008 Early Stages of Calcium Phosphate Formation 1531

Fig. 3. Scanning electron microscopy images, at lower and higher magnifications, of the surfaces of H12 glass disks reacted in aqueous phosphate
solution for 0.5 h (a), (b), 1 h (c), (d), and 3 h (e), (f).

Fig. 4. Scanning electron microscopy images of the surface of H12 glass disk reacted in aqueous phosphate solution for 3 h, showing layered-structure,
plate-like crystals identified by X-ray diffraction as brushite, CaHPO4  2H2O.
1532 Journal of the American Ceramic Society—Li et al.
Fig. 5. Scanning electron microscopy images of the surfaces of H12 glass disks reacted in aqueous phosphate solution for 5 h (a), (b) and 1 day (c), (d).
water.20–23 On the other hand, for the bioactive glass–ceramic
A–W, a calcium phosphate layer formed in SBF but not in
deionized water.24 For a given material (bioactive glass or glass–
ceramic), a minimum concentration of phosphate ions should be
reached near the surface for a calcium phosphate layer to pre-
cipitate. This concentration should depend on the P2O5 content
of the starting material and the duration of the treatment in the
liquid.21
For glasses in the system CaO–SiO2, the addition of a small
quantity of Al2O3 reduced the bioactivity, and the addition of
more than 1.7 mol% Al2O3 was found to suppress the bioac-
tivity, by suppressing the formation of a calcium phosphate
layer on the glass surface in an SBF.25 The present H12 boro-
silicate glass contained 2.5 mol% Al2O3, but still showed the
ability to readily form a calcium phosphate layer on its surface.
The concentration of Al2O3 required to suppress the bioactivity
is expected to depend on the composition (or chemical durabil-
ity) of the glass. Because of their lower durability, some borate
glasses react rapidly in a aqueous phosphate solution,9,10 so the
amount of Al2O3 required to suppress their bioactivity is
expected to be higher. The H12 glass used in this work,
although referred to technically as a borosilicate glass, is essen-
tially a borate glass (40 mol% B2O3) containing a small amount
of SiO2 (7.5 mol%). Further work is required to systematically
study the effect of Al2O3 concentration on the bioactivity of
borate-based glass compositions.
IV. Conclusion
The early-stage formation of calcium phosphate during the con-
version of a bioactive borosilicate glass (designated H12) was
found to be different from that for both silicate 45S5 bioactive
glass and borate bioactive glass. When disks of H12 glass were
reacted in 0.25M K2HPO4 solution with pH 5 7.0 at 371C, a
crystalline calcium hydrogen phosphate dihydrate (brushite),
Vol. 91, No. 5
CaHPO4  2H2O, with a coarse plate-like morphology, was
observed to form on the glass surface at short reaction times
(3–5 h). In comparison, the early-stage conversion product is
an SiO2-rich gel layer followed by an ACP layer for 45S5 glass,
and an ACP layer for borate glass. For longer reaction times
(4B5 h), the conversion product was essentially HA, with
small, needle-like crystals, similar to the later-stage product
observed for several bioactive glasses. Possible reasons for the
early-stage formation of CaHPO4  2H2O on H12 glass may
include kinetic factors, the influence of the glass composition,
or a combination of both.
References
1
L. L. Hench, R. J. Splinter, W. C. Allen, and T. K. Greenlee Jr., ‘‘Bonding
Mechanisms at the Interface of Ceramic Prosthetic Materials,’’ J. Biomed. Mater.
Res. Symp., 2, 117–41 (1971).
2
L. L. Hench and H. A. Paschall, ‘‘Direct Chemical Bond of Bioactive Glass-
Ceramic Materials to Bone and Muscle,’’ J. Biomed. Mater. Res. Symp., 4, 25–42
(1973).
3
L. L. Hench and J. Wilson, ‘‘Surface-Active Biomaterials,’’ Science, 226, 630–6
(1984).
4
L. L. Hench, ‘‘Bioceramics: From Concept to Clinic,’’ J. Am. Ceram. Soc., 74
[7] 1487–510 (1991).
5
L. L. Hench, ‘‘Bioceramics,’’ J. Am. Ceram. Soc., 81 [7] 1705–28 (1998).
6
M. N. Rahaman, R. F. Brown, B. S. Bal, and D. E. Day, ‘‘Bioactive Glasses for
Nonbearing Applications in Total Joint Replacement,’’ Semin. Arthroplasty, 17,
102–12 (2006).
7
Q. Fu, M. N. Rahaman, B. S. Bal, W. Huang, and D. E. Day, ‘‘Preparation
and Bioactive Characteristics of a Porous 13–93 Glass, and Fabrication into the
Articulating Surface of a Proximal Tibia,’’ J. Biomed. Mater. Res., 82A, 222–9
(2007).
8
N. W. Marion, W. Liang, G. C. Reilly, D. E. Day, M. N. Rahaman, and J. J.
Mao, ‘‘Bioactive Borate Glass Supports the Osteogenic Potential of Human Me-
senchymal Stem Cells,’’ Mech. Adv. Mater. Struct., 12, 239–46 (2005).
9
W. Huang, D. E. Day, K. Kittiratanapiboon, and M. N. Rahaman, ‘‘Kinetics
and Mechanism of the Conversion of Silicate (45S5), Borate, and Borosilicate
Glasses to Hydroxyapatite in Dilute Phosphate Solutions,’’ J. Mater. Sci. Mater.
Med., 17, 583–96 (2006).
10
W. Huang, M. N. Rahaman, D. E. Day, and Y. Li, ‘‘Mechanisms for Con-
verting Bioactive Silicate, Borate, and Borosilicate Glasses to Hydroxyapatite in
May 2008 Early Stages of Calcium Phosphate Formation
Dilute Phosphate Solution,’’ Phys. Chem. Glass.: Europ. J. Glass Sci. Technol. Part
B, 47 [6] 647–58 (2006).
11
H. K. Teitelbaum, ‘‘Borate Glasses for Bioactive Coatings on Titanium’’; M.S.
Thesis, University of Missouri-Rolla, 2002.
12
L. Peddi, ‘‘Bioactive Borate Glass Coatings for Titanium and its Alloys’’; M.S.
Thesis, University of Missouri-Rolla, 2005.
13
Y. Li, M. N. Rahaman, B. S. Bal, and D. E. Day, ‘‘Conversion of Bioactive
Borosilicate Glass to Multilayered Hydroxyapatite in Dilute Phosphate Solution,’’
J. Am. Ceram. Soc., 90 [12] 3804–10 (2007).
14
A. E. Clark and L. L. Hench, ‘‘Early Stages of Calcium-Phosphate Layer
Formation in Bioglass,’’ J. Non Cryst. Solids, 113, 195–202 (1989).
15
M. R. Filgueiras, G. LaTorre G, and L. L. Hench, ‘‘Solution Effects on the
Surface Reaction of a Bioactive Glass,’’ J Biomed. Mater. Res., 27, 445–53 (1993).
16
D. E. Day, J. E. White, R. F. Brown, and K. D. McMenamin, ‘‘Transfor-
mation of Borate Glasses into Biologically Useful Materials,’’ Glass Technol., 44
[2] 75–81 (2003).
17
G. H. Nancollas and J. Zhang, ‘‘Formation and Dissolution Mechanisms of
Calcium Phosphates in Aqueous Systems’’; pp. 73–81 in Hydroxyapatite and
Related Materials, Edited by P. W. Brown, and B. Constantz. CRC Press, Boca
Raton, FL, 1994.
1533
18
G. W. Van Loon and A. I. Omoike, ‘‘Removal of Phosphorus and
Organic Matter by Alum During Wastewater Treatment,’’ Water Res., 33,
3617–27 (1999).
19
T. Kokubo and H. Takadama, ‘‘How Useful is SBF in Predicting in vivo Bone
Bioactivity,’’ Biomaterials, 27, 2907–15 (2007).
20
C. G. Pantano Jr., A. E. Clark Jr., and L. L. Hench, ‘‘Multilayer Corrosion
Films on Bioglass Surfaces,’’ J. Am. Ceram. Soc., 57 [9] 412–3 (1974).
21
A. E. Clark Jr., C. G. Pantano Jr., and L. L. Hench, ‘‘Auger Spectroscopic
Analysis of Bioglass Corrosion Films,’’ J. Am. Ceram. Soc., 59 [1–2] 37–9 (1976).
22
M. Ogino, F. Ohuchi, and L. L. Hench, ‘‘Compositional Dependence of the
Formation of Calcium Phosphate Films on Bioglass,’’ J. Biomed. Mater. Res., 14,
55–64 (1980).
23
M. Ogino and L. L. Hench, ‘‘Formation of Calcium Phosphate Films on Sil-
icate Glasses,’’ J. Non-Cryst. Solids, 38/39, 673–8 (1980).
24
T. Kokubo, H. Kushitani, S. Sakka, T. Kitsugi, and T. Yamamuro, ‘‘Solu-
tions Able to Reproduce in vivo Surface-Structure Changes in Bioactive Glass–
Ceramic A-W,’’ J. Biomed. Mater. Res., 24, 721–34 (1990).
25
C. Ohtsuki, T. Kokubo, and T. Yamamuro, ‘‘Compositional Dependence of
Bioactivity of Glasses in the System CaO–SiO2–Al2O3: Its In Vitro Evaluation,’’
J. Mater. Sci.: Mater. Med., 3, 119–25 (1992). &