1. Investigating the effect of TiO2 on the structure and biocompatibility of
bioactive glass
Lana M. Placek,1
Timothy J. Keenan,1
Yiming Li,1
Chokchai Yatongchai,1
Dimple Pradhan,1
Daniel Boyd,2
Nathan P. Mellott,1
Anthony W. Wren1
1
Inamori School of Engineering, Alfred University, Alfred, New York
2
Department of Applied Oral Sciences, Dalhousie University, Halifax, Nova Scotia, Canada
Received 7 May 2015; revised 4 August 2015; accepted 23 August 2015
Published online 7 September 2015 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/jbm.b.33521
Abstract: Titanium (Ti41
) containing materials have been
widely used in medical applications due to its associated bio-
activity in vivo. This study investigates the replacement of
Si41
with Ti41
within the system SiO2-Na2O-CaO-P2O5 to deter-
mine its influence on glass structure. This strategy was con-
ducted in order to control the glass solubility to further
improve the cellular response. Ti41
incorporation was found
to have little influence on the glass transition temperature
(Tg 5 520 6 88C) and magic angle spinning-nuclear magnetic
resonance (MAS-NMR) shifts (280 ppm) up to additions of 18
wt %. However, at 30 wt % the Tg increased to 6008C and MAS-
NMR spectra shifted to 288 ppm. There was also an associ-
ated reduction in glass solubility as a function of Ti41
incorpo-
ration as determined by inductively coupled plasma optical
emission spectroscopy where Si41
(1649–44 mg/L) and Na1
(892–36 mg/L) levels greatly reduced while Ca21
(3–5 mg/L)
and PO32
4 (2–7 mg/L) levels remained relatively unchanged.
MC3T3 osteoblasts were used for cell culture testing and it
was determined that the Ti41
glasses increased cell viability
and also facilitated greater osteoblast adhesion and prolifera-
tion to the glass surface compared to the control glass. VC 2015
Wiley Periodicals, Inc. J Biomed Mater Res Part B: Appl Bio-
mater, 104B: 1703–1712, 2016.
Key Words: bioactive glass, titanium, ion release, cell culture,
solubility
How to cite this article: Placek LM, Keenan TJ, Li Y, Yatongchai C, Pradhan D, Boyd D, Mellott NP, Wren AW. 2016.
Investigating the effect of TiO2 on the structure and biocompatibility of bioactive glass. J Biomed Mater Res Part B
2016:104B:1703–1712.
INTRODUCTION
Bioglass is regarded as one of the most widely studied and
commercially successful bioceramics applied in orthopedics
today. This material was originally developed in 1969 by
Prof. Larry Hench at the University of Florida and the sub-
sequent class of materials that stemmed from this research,
bioactive glasses, has been used to formulate numerous
glass-based biomaterials today. Early studies on the original
45S5 Bioglass formulation determined that a permanent sta-
ble bond to bone could be established in animal models,
which has led to Bioglass being marketed and applied in
the medical field as particulates and pastes.1,2
Additional
bioactive glass-based materials include glass-ceramic scaf-
folds for orthopedic applications,3–5
glass polyalkenoate
bone cements,6–8
and glass microspheres for cancer treat-
ment.9–11
Since their inception, bioactive glasses have been
applied and investigated for hard tissue repair1,2,12
as they
stimulate osteogenesis in in vitro models.1,2
Bioactive
glasses are characterized by their ability to promote healing
within the body due to the dissolution of the glass surface
and ion exchange upon exposure to physiological medium
or body fluids.13,14
The specific mechanism includes soluble
Si41
being released into the surrounding medium in the
form of silicic acid due to ion exchange with H1
and
H3O1
.12
This subsequently results in the precipitation of an
amorphous calcium phosphate layer which crystallizes to a
carbonated hydroxyapatite layer (HCA).14,15
Additionally, it
has been established that ionic dissolution products from
45S5 Bioglass and other silicate-based glasses can stimulate
angiogenesis and expression of several genes of osteoblasts,
which suggests their applicability may supersede primarily
osseous tissue.15–17
This study aims to determine any changes in the solubil-
ity and cytocompatibility of a novel SiO2-Na2O-CaO-P2O5
glass series, where titanium Ti41
is substituted for Si41
in 6
wt % increments. Ti41
has previously been incorporated
into numerous medical materials as it is known to promote
apatite formation on the materials surfaces when in contact
with physiological fluids, that is, Ti-6Al-4V,18–21
Ti-gels,14
coatings,22
and glasses.23–25
Cytocompatibility studies on
the effect of Ti41
in cells such as monocytes/macrophages26
and leukocytes have been conducted and resulted in no sig-
nificant changes in cell viability.27
Additional studies in
monocyte derived dendritic cells (DCs) resulted in increased
Correspondence to: A. W. Wren; e-mail: wren@alfred.edu
VC 2015 WILEY PERIODICALS, INC. 1703

2. T-lymphocyte activity.28
Further testing in cell lines more
relevant to bone (osteoblast like ROS) resulted in increased
expression of ALP, OPN, and osteonectin, thereby indicating
their role as promoters of osteoblast differentiation.29,30
Ion
concentration from these studies cite that 0.1–10 mg/L Ti41
levels did not significantly alter cell viability, however,
decreases were evident when the concentration reached
20 mg/L.31
In relation to incorporating TiO2 into bioactive glasses,
the authors have completed studies that investigated the
structure and solubility of TiO2 doped silicate glasses. These
studies on CaO-SrO-ZnO-SiO2/TiO2 glasses determined Ti41
to act predominantly as a network modifier as its replace-
ment for Si41
greatly increased the concentration of non-
bridging oxygen (NBO) species within the glass.32
In rela-
tion to glass structure it has previously been determined
that the concentration of NBO (Si-O-NBO) species within a
glass is directly related to its solubility.33
NBO groups, cre-
ated by the inclusion of alkali/alkali earth cations, facilitate
the ion exchange process and promote the rate of silica dis-
solution further increasing bioactivity.33
Cations play a com-
plex structural role in oxide glasses, allowing them to
influence the physical and chemical properties of the result-
ing glass.32,34
Cations are usually classified as network for-
mers, which form the basic structure of the glass network,
or network modifiers, which de-polymerize the silicate glass
network and ensure charge compensation for non-
tetravalent network forming elements. Glass properties such
as ion release and pH may be strongly influenced depending
on the structural role of its cations,34
specifically, with
respect to this study, the effect of increasing TiO2 concentra-
tion will be investigated as Ti41
can assume a network
modifying or network forming role in a glass. Previous stud-
ies by the authors on the solubility of Ti41
substituted sili-
cates glasses (SiO2-TiO2-Na2O-CaO) was found to greatly
reduce the rate of Ca21
release as the concentration of Ti41
increased within the glass which had a positive influence on
the viability of L929 fibroblast cells.35
With respect to this
study, a TiO2 substituted bioactive glass (SiO2-Na2O-CaO-
P2O5) series was produced by the traditional melt-quench
method and was investigated with respect the glass struc-
ture, solubility, (ion release and pH) and the subsequent
effect on MC3T3 osteoblast cells. Controlling the solubility
of bioactive materials surfaces, such as glass, can greatly
increase the cellular response, in particular growth and
adhesion. For this study magic angle spinning-nuclear mag-
netic resonance (MAS-NMR) was used to investigate the
structure of the glasses while the glass solubility was deter-
mined using inductively coupled plasma optical emission
spectroscopy (ICP-OES). To conclude, cell viability and adhe-
sion studies were conducted on each of the glass composi-
tions using appropriate cell lines such as MC3T3
osteoblasts.
MATERIALS AND METHODS
Glass synthesis
Glass powder production. Five glass compositions (BG-1,
BG-2, BG-3, BG-4, and BG-5) were formulated for this study
with the principal aim being to investigate structure and
solubility changes with the substitution of TiO2 for SiO2
within the glass composition (42SiO2224Na2O-21CaO-
13P2O5 wt %) with Ti41
replacing Si41
at 6 wt % (BG-1)
increments up to 30 wt % (BG-5). A control glass (Control)
was also formulated which did not contain TiO2. Glasses
were prepared by weighing out appropriate amounts of ana-
lytical grade reagents (SiO2, TiO2, CaCO3, Na2CO3,
NH4H2PO4) and ball milling (1 h). The powdered mixes
were fired (15008C, 1 h) in platinum crucibles and shock
quenched in water. The resulting frits were dried, ground,
and sieved to retrieve glass powders ranging from 90 to
710 lm. To ensure consistent surface area/particle size,
each glass was ultrasonically cleaned in isopropyl alcohol
for 30 min and dried in an oven at 378C overnight.
Material characterization
X-ray diffraction. Diffraction patterns were collected using
a Siemens D5000 X-ray Diffraction (XRD) Unit (Bruker AXS,
WI). Glass powder samples were packed into standard
stainless steel sample holders. A generator voltage of 40 kV
and a tube current of 30 mA were employed. Diffractograms
were collected in the range 108 < 2h < 808, at a scan step
size 0.028 and a step time of 10 s. Any crystalline phases
present were identified using Joint Committee for Powder
Diffraction Studies (JCPDS) standard diffraction patterns.
Differential thermal analysis. A combined differential ther-
mal analyzer-thermal gravimetric analyzer (DTA-TGA) (Stan-
ton Redcroft STA 1640, Rheometric Scientific, Epsom, UK)
was used to measure the glass transition temperature (Tg)
for all glasses. A heating rate of 108C min21
was employed
using an alumina crucible where a matched alumina cruci-
ble was used as a reference. Sample measurements were
carried out every six seconds between 30 and 13008C.
MAS-NMR. 29
Si MAS NMR spectra were recorded using a 14
T (tesla) Bruker Avance III wide-bore FT-NMR spectrometer
(Billerica, MA), equipped with a double broadband tunable
triple resonance HXY CP-MAS probe. The glass samples
were placed in a zirconia sample rotor with a diameter of
4 mm. The sample spinning speed at the magic angle to the
external magnetic field was 10 kHz. 29
Si MAS NMR spectra
were acquired at 300 K with the transmitter set to 119.26
MHz (2100 ppm) with a 3.0 us pulse length (pulse angle,
p/2), 120-s recycle delays, where the signals from 640
scans were accumulated for each glass powder. 29
Si NMR
chemical shifts are reported in ppm, with trimethylsilylpro-
pionate (TMSP) as the external reference (0 ppm). Data
were processed using a 25 Hz Gaussian apodization function
followed by baseline correction and were deconvoluted
using PeakFit version 4.12.
Scanning electron microscopy/energy dispersive X-ray
spectroscopy. A Quanta 200F Environmental Scanning Elec-
tron Microscope (SEM) was used to image the samples
under a vacuum at a pressure of 0.90 torr. The electron
beam was used at an accelerating voltage of 20 kV and a
1704 PLACEK ET AL. INVESTIGATING THE EFFECT OF TIO2

3. spot size of 4.0. Energy dispersive X-ray (EDX) spectroscopy
was carried out using an FEI EDAX system equipped with a
silicon-drift detector.
Surface area analysis. In order to determine the surface
area of the glass particles the Advanced Surface Area and
Porosimetry, ASAP 2010 System analyzer (Micrometrics
Instrument Corporation, Norcross) was employed. Approxi-
mately 60 mg of each glass was used to calculate the spe-
cific surface areas using the Brunauer-Emmett-Teller (BET)
method.
Glass solubility
Ion release profiles (ICP). Each glass (Control, BG-1, BG-2,
BG-3, BG-4, BG-5, n 5 3 where n is the number of samples)
was immersed in sterile deionized H2O for 1, 7, and 14
days. Glass particulates (1 m2
surface area) was submerged
in 10 mL of deionized H2O and rotated on an oscillating
platform at 378C. The ion release profile of each specimen
was measured using ICP–atomic emission spectroscopy
(ICP–AES) on a Perkin-Elmer Optima 5300UV (Perkin Elmer,
MA,). ICP–AES calibration standards for Si, Ti, Na, Ca, and P
ions were prepared from a stock solution on a gravimetric
basis. Three target calibration standards were prepared for
each ion and de-ionized water was used as a control.
pH analysis. Changes in solution pH were monitored using
a Corning 430 pH meter (Fisher Scientific, Pittsburgh, PA).
Sample solutions were prepared by exposing glass particu-
late samples (1 m2
surface area) in 10 mL of deionized H2O.
Measurements were recorded at 1, 7, and 14 days. Sterile
deionized water was used as a control and was measured at
each time period.
Biocompatibility testing
Sample preparation. Glass samples (Control, BG-3, and BG-
5) were analyzed for cytocompatibility using both MTT test-
ing (glass particles) and cell adhesion studies (polished
glass plates). For cell viability analysis glass particulates
were incubated (1 m2
) in sterile deionized water for 1, 7,
and 14 days. Liquid extracts were then removed and used
for MTT assay testing. For cell adhesion studies, glass plates
(6 mm 3 1.5) were used. Each sample was initially auto-
claved and then washed with 10 volumes of a mixture of
50% acetone and 50% ethanol once for 30 min. They were
then sonicated in 70% ethanol and rinsed three times with
100% ethanol for 30 min each in an orbital shaker at room
temperature. They were then air-dried for 15 min. Glass
plates were buffered for 72 h by immersing them in PBS
and shaking at 50 rpm, at 378C. The PBS solution was
exchanged every 8 h. Alterations in the pH of the buffer was
monitored until the pH was 7.6–7.85. Furthering buffering
was carried out in Dulbecco’s modified eagle medium for
24 h (n 5 3/sample). The final pH of buffering medium was
7.4–7.5. Buffered glasses were vacuum-dried for 15 min
and heated to 1208C for 2 h for further sterilization.
MTT assay. The established cell line MC3T3-E1 osteoblasts
(ATCC CRL-2593) was employed for biocompatibility analy-
sis. Cells were maintained on a regular feeding regime in a
cell culture incubator at 378C/5% CO2/95% air atmosphere.
Cells were seeded into 96 well plates at a density of 10,000
cells per well and incubated for 24 h prior to testing with
glass liquid extracts. The culture media used was Minimum
Essential Medium Alpha Media (Fisher Scientific, PA) sup-
plemented with 10% fetal bovine serum (Fisher Scientific,
PA) and 1% (2 mM) L-glutamine (Fisher Scientific, PA). The
cytotoxicity of glass liquid extracts (Control, BG-3 and BG-5)
was evaluated using the Methyl Tetrazolium (MTT) assay in
96 well plates. Liquid extracts (10 mL) were added into
wells containing osteoblast cells in culture medium (100
mL) in triplicate. Each of the prepared plates was incubated
for 24 h at 378C/5% CO2. The MTT assay was then added in
an amount equal to 10% of the culture medium volume/
well. The cultures were then reincubated for a further 3 h
(378C/5% CO2). Next, the cultures were removed from the
incubator and the resultant formazan crystals were dis-
solved by adding an amount of MTT Solubilization Solution
(10% Triton x-100 in Acidic Isopropanol (0.1N HCI)) equal
to the original culture medium volume. Once the crystals
were fully dissolved, the absorbance was measured at a
wavelength of 570 nm. Aliquots (10 mL) of sterile media
were used as a control, and cells were assumed to have
metabolic activities of 100%.
Osteoblast adhesion procedure. The mouse calvaria-
derived MC3T3-E1 preosteoblast cells (ATCC) were cultured
in Minimum Essential Medium Alpha Media supplement
with 10% fetal bovine serum (Fisher Scientific, PA). Cells
were maintained on a regular feeding regime in a cell cul-
ture incubator at 378C/5% CO2/95% air atmosphere. After
28–48 h incubation, media was removed and 5 mL trypsin
was added to the culture flask. The cells were left to detach
for 20 min, after this time, trypsin was removed and cells
were resuspended in culture media. The number of cells
was calculated to 10,000 cells per 1 mL media. The condi-
tioned glass plates were placed in each well (n 5 3/sample)
where 1 mL cell/media solution was seeded onto the sur-
face of the glass plates and incubated for 24 h.
Sample preparation for SEM. Samples were fixed with 4%
(w/v) paraformaldehyde in 13 PBS buffer for 30 min and
then post-fixed with 1% osmium tetroxide in distilled water
for 1 h. Samples were dehydrated with a series of graded
ethanol washes (50/60/70/80/90/100% DI water). Sam-
ples were immersed in hexamethyldislizane for 5 min and
then transferred to a desiccator for 30 min. The glass plates
were then coated in gold and sample imaging was carried
out using an FEI Co. Quanta 200F Environmental SEM
equipped with an EDAX Genesis Energy-Dispersive
Spectrometer.
Statistical analysis. One-way analysis of variance
(ANOVA) was employed to compare the experimental mate-
rials in relation to (1) differences cell viability as a function
of glass composition, and (2) the effect each glass on cell
ORIGINAL RESEARCH REPORT
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH B: APPLIED BIOMATERIALS | NOV 2016 VOL 104B, ISSUE 8 1705

4. viability as a function of maturation in aqueous media. Com-
parison of relevant means was performed using the post
hoc Bonferroni test. Differences between groups was
deemed significant when p 0.05.
RESULTS
Influence of Ti41
on glass structure
Glasses in this study were formulated using the melt-
quench method where 6 wt % TiO2 is substituted in incre-
mental conditions for SiO2 up to 30 wt % TiO2. Initial char-
acterization included XRD and is presented in Figure 1. The
Control and BG-1 were found to be fully amorphous while
BG-2 and BG-3 presented some minor crystal peaks at 228,
248, 318, 338, 418, 4682h which correspond to Na2Ca4(PO)2-
SiO4 (00–033-1229). BG-4 and BG-5 were also found to be
fully amorphous with a BG-5 presenting a primary amor-
phous hump at approximately 328 2h while a secondary
hump was evident at a lower position of approximately
1882h. Additionally, the surface area of the glass particles
was relatively consistent and ranged from 0.02 to 0.06m2
/g
(Table I). DTA is presented in Figure 2 and was conducted
to determine any changes in the glass transition tempera-
ture (Tg) as a function of TiO2 concentration. From Figure 2
it is evident Control has a Tg 5 5208C. The addition of 6, 12,
and 18 wt % TiO2 resulted in an increase of 108C to
5278C, 5288C, and 5248C, respectively. Higher additions of
TiO2 (BG-4, BG-5) resulted in Tg increases to 5938C and
6008C, respectively. Additionally, higher additions of TiO2
resulted in increasing the crystallization transitions where
Control first crystallization peaks is present at 6028C while
the first crystallization peak attributed to BG-5 is observed
at 6528C. The ratio of Si/Ti decreased from 6.0 (6 wt % Ti,
BG-1) to 0.6 (30 wt % Ti, BG-5) and this trend is also pre-
sented in Figure 2 (Table II).
Glass composition, particle size and morphology were
evaluated using SEM and energy dispersive X-ray (EDS)
analysis. EDS analysis is presented in Figure 3 for Control,
BG-1, BG-3, and BG-5 and confirms the elemental composi-
tion where Si41
, Ca21
, PO32
4 , and Na1
are present in Con-
trol, and the same elements in addition to Ti41
are present
in the modified glass series. Figure 4 shows SEM images of
each glass where it is evident that the large glass particles
are free of smaller particles. The morphology of each parti-
cle is relatively similar and ranges from approximately 500
to 700 lm in diameter.
MAS-NMR was conducted to determine any changes in
glass structure as a function of Ti41
substitution. Figure 5
presents MAS-NMR spectra of select glasses including Con-
trol, BG-1, BG-3, and BG-5. Figure 5(a) presents Control
which shows a main peak centered at 286.0 ppm. The main
peak was de-convoluted to reveal a peak at 279.3, 286.0,
295.4, and 2107.4 ppm. Figure 5(b) presents BG-1 which
presents a main peak at 283.0 ppm and de-convoluted
peaks at 280.9, 287.1, and 2106.6 ppm. Figure 5(c)
presents the NMR spectra of BG-3 which has a main peak
centered at 280.3 ppm, where de-convolution produced
peaks centered at 275.9, 280.3, 289.4, and 289.5 ppm.
Figure 5(d) present the spectra of BG-5 which shifted in a
FIGURE 1. X-ray diffraction of Control and Ti-series.
FIGURE 2. Relationship between the glass transition temperature (Tg)
and the Si/Ti ratio throughout the Ti-glass series.
TABLE I. Surface Area of Prepared and Cleaned Glass
Particles
Control BG1 BG2 BG3 BG4 BG5
Surface
area (m2
/g)
0.054 0.056 0.046 0.038 0.023 0.017
TABLE II. Summary of Ti41
Glass Series Characterization
XRD Tg (8C)
NMR
(ppm)
Si:Ti
Ratio Q-Structure
Control Amorphous 520 286 – Q2
BG1 Amorphous 527 283 6.0 Q2
BG2 Partial Crys. 528 – 2.6 Q2
BG3 Partial Crys. 524 280 1.5 Q2
BG4 Amorphous 593 – 0.9 Q2
BG5 Amorphous 600 288 0.6 Q2/Q3
1706 PLACEK ET AL. INVESTIGATING THE EFFECT OF TIO2

5. more negative direction, with the main peak centered at
288.0 ppm. De-convolution revealed peaks at 287.1,
297.1, and 2105.0 ppm.
Influence of Ti41
on glass solubility and bioactivity
Ion release profiles for Silica (Si41
) and Sodium (Na1
) are
presented in Figure 6. Ion release was measured over 1, 7,
and 14 days. Figure 7(a) presents Si41
release which was
highest for Control which was found to increase with
respect to time, where it increased from 90 to 1649 mg/L
over 1–14 days. Si41
levels reduced with respect to compo-
sition (i.e., the replacement of Si41
with Ti41
), where after
14 days incubation Si41
levels were, BG-1 (344 mg/L), BG-2
(243 mg/L), BG-3 (83 mg/L), BG-4 (58 mg/L), and BG-5
(44 mg/L), and each followed a similar trend to Control
where Si41
release increased over time. Figure 6(b)
presents Na1
release which presented similar trends where
Control experienced the highest release rates which peaked
at 14 days at 892 mg/L. Na1
levels peaked at 14 days, how-
ever, they were reduced as a consequence of modifying the
glass composition, BG-1 (224 mg/L), BG-2 (177 mg/L), BG-3
(90 mg/L), BG-4 (33 mg/L), and BG-5 (36 mg/L), and each
also followed a similar trend to Control where Na1
release
increased over time. Ion release profiles are presented in
Figure 7 for Calcium (Ca21
) and Phosphorus (P) over 1, 7,
and 14 days. With respect to Ca21
release, each material
reduced with respect to time. Ca21
release attributed to
Control was highest at 1 day at 5 mg/L in addition BG-1 at
6 mg/L. However, with increasing Ti41
concentration in the
glass, there was an overall reduction in Ca21
release, BG-2
(4 mg/L), BG-3 (2 mg/L), BG-4 (2 mg/L), and BG-5 (2 mg/
L) at 1 day. Release rates at 14 days were 0 mg/L for Con-
trol and ranged from 0.2 to 0.5 mg/L, with BG-5 attaining
the highest Ca21
release at 0.5 mg/L. PO32
4 release did not
present any visible trend with respect to time or Ti41
con-
centration. PO32
4 release ranged from 2 to 7 mg/L and was
highest for BG-2 at 14 days at 7 mg/L. Ti41
release was not
determined for any of the glass compositions at 1, 7, or 14
days incubation. pH changes were monitored and recorded
after each time period, that is, 1, 7, 14 days, and are pre-
sented in Figure 8. Control produced the highest pH over 1–
14 days (10.9–12.1) and increased with respect to time. As
the Ti41
concentration increased in the modified glasses the
pH was found to decrease, however, each glass did produce
an increase in pH with respect to time. At 14 days pH val-
ues were recorded for each material as BG-1 (10.9), BG-2
(11.5), BG-3 (11.1), BG-4 (10.3), and BG-5 (10.4). The exis-
tence of low level crystallinity in BG-3 was not found to sig-
nificantly influence the solubility profiles in comparison to
each of the amorphous counterparts.
Biocompatibility testing was conducted using cell viabil-
ity (MTT assay) and cell adhesion studies. Cell viability stud-
ies were subject to statistical comparison in order to
compare differences cell viability as a function of glass com-
position, and to determine the influence each glass has on
cell viability as a function of maturation in aqueous media
(statistics presented in discussion section). Figure 9 shows
that Control presented a viability of 75% after 1 day com-
pared to the growing cell population (100%). These levels
increased to 100% after 7 days and further increased to
115% after 14 days. Regarding the Ti41
containing glasses,
BG-1 cell viability increased from 93%, 104% and 115%
FIGURE 3. Energy dispersive X-ray analysis of Control and Ti-series.
FIGURE 4. Scanning electron microscopy of Control and Ti-series.
ORIGINAL RESEARCH REPORT
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH B: APPLIED BIOMATERIALS | NOV 2016 VOL 104B, ISSUE 8 1707

6. after 1, 7, and 14 days, respectively. BG-3 presented similar
levels at 104%, 95% and 110% after 1, 7, and 14 days. BG-
5 presented higher levels of cell viability of 121% after 1
day, 132% after 7 days, and 103% after 14 days. Cell adhe-
sion studies are presented in Figures 10 and 11, and shows
that osteoblast did not adhere on the surface of Control
FIGURE 5. MAS-NMR of a) Control, b) BG1, c) BG3, and d) BG5.
FIGURE 6. Ion release of Control and Ti-series, specifically a) Silica and b) Sodium.
1708 PLACEK ET AL. INVESTIGATING THE EFFECT OF TIO2

7. samples. However, the Ti41
containing glasses that were
tested (BG-3 and BG-5) presented osteoblast adhesion on
the glass surfaces after 24 h incubation in culture medium.
DISCUSSION
Influence of Ti41
on glass structure
This study was conducted in order to determine the effect
that substituting SiO2 with TiO2 has on the structure, solu-
bility and biocompatibility of bioactive glass. Initial charac-
terization determined that crystal peaks (Figure 1) were
identified in BG-2 and BG-3, which is likely an effect of glass
processing. DTA determined that the Tg of the glasses con-
taining 6, 12 and 18 wt % TiO2 showed relatively little devi-
ation compared to the Control. This is likely due to Ti41
structural role in the glass where, up to BG-3 (18 wt %) it
may act predominantly as a network modifier. This is fur-
ther evidenced by the MAS-NMR spectra (Figure 5) where,
with the increase in Ti41
concentration, the NMR spectra
shifts from 286.0 ppm (Control) to 283.0 ppm (BG-1, 6 wt
% Ti41
) and finally 280.3 ppm (BG-3, 18 wt % Ti41
). How-
ever, with the addition of 30 wt % Ti41
(BG-5), a spectral
shift was observed to occur to 288.0 ppm, suggesting a
change in the structural role of Ti41
. An established method
of representing the structural arrangement of a glass can be
represented by Qn
units, where Q represents the Si tetrahe-
dral unit and n the number of bridging oxygens (BO); where
n ranges between 0 and 4. Si41
is the central tetrahedral
atom which ranges from Q0
(orthosilicates) to Q4
(tectosili-
cates) and Q1
, Q2
, and Q3
structures representing intermedi-
ate silicates containing modifying oxides.36
Previous studies
on silicate glasses have identified regions where the chemi-
cal shift represents structural changes around the Si41
atom
which lie in the region of 260 to 2120 ppm for SiO4 tetra-
hedra. Previous NMR studies by Galliano et al.37
and Haya-
kawa et al.38
on silicate melts suggest the presence of Q1
,
Q2
, Q3
, and Q4
species at 278, 285, 295, and 2105 ppm,
respectively. The de-convolution revealed that the Control
glass has a greater distribution of Q-species than the BG-
Series as spectral resolution ranges from 279.3 to 2107.4
ppm, suggesting each of these Q-species is represented.
With respect to BG-1 and BG-3 it is evident that as Ti41
is
incorporated into the glass structure, a shift in a positive
direction to 283.0 and 280.3 ppm is presented. BG-1 spec-
tral shift ranges from 280.9 to 2106.6 ppm, similar to
FIGURE 7. Ion release of Control and Ti-series, specifically a) Calcium and b) Phosphate.
FIGURE 8. pH changes of Control and Ti-series recorded over 1, 7,
and 14 days.
FIGURE 9. MTT assay of bioactive glass series over 1, 7, and 14 days
using MC3T3 osteoblast cells.
ORIGINAL RESEARCH REPORT
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH B: APPLIED BIOMATERIALS | NOV 2016 VOL 104B, ISSUE 8 1709

8. Control, while BG-3 ranges from 275.0 to 289.5 ppm which
predominantly represents Q1
-Q3
structure. This shift in ppm
is representative of a reduction in BO content and an
increase in the concentration of NBO, a characteristic that
has been previously identified in CaO-SrO-ZnO-SiO2/TiO2
using X-ray photoelectron spectroscopy.32
The NMR shift
presented for BG-5 [Figure 5(d)] differs from each of the
other spectra with a center at 288 ppm and de-convoluted
peaks ranging from 287.1 to 2105.0 ppm. This suggests a
higher concentration of BO with the addition of 30 wt %
TiO2. This may be due to the ratio of Si:Ti within the glass.
Previous studies on alumino-silicate glasses have deter-
mined that Al31
can change state from a network modifier
to a network former (as Al31
is considered a network inter-
mediate, similar to Ti41
) when the ratio of Si:Al exceeds 1.
In this case, the Si:Ti ratio decreases from 6.0 (BG-1) to 0.9
(BG-4) and 0.6 (BG-5) with the addition of up to 30 wt %
Ti41
. As the ratio of Si:Ti decreases, and approaches 1, there
are structural changes occurring within the glass as evident
by DTA and MAS-NMR data. With respect to BG-4 and BG-5
there is a significant increase in Tg and an associated shift
in the negative direction when measured by MAS-NMR for
BG-5. This is in contrast to previous studies on TiO2 doped
silicate-based glasses studied by the authors where the Tg
measurements were found to consistently reduce with
increased Ti41
concentration.32,35
The relatively insignificant
change in Tg associated with BG-2 and BG-3 may be due to
the partial crystallinity (Figure 1) where the crystal struc-
tures produced may hinder molecular movement within the
glass as these crystal structures are not soluble at this tem-
perature (520–5308C). This may result in an expected
reduction in Tg (Figure 2) if the NBO concentration is
increasing. This may be the reason the Tg does not correlate
with MAS-NMR spectra where with an increase in TiO2 con-
centration, there is an associated de-polymerization of the
Si41
network. It can be observed with the amorphous BG-4
and BG-5 (Figure 1) that there is an increase in Tg which
correlates MAS-NMR data for BG-5 particularly.
Influence of Ti41
on glass solubility and bioactivity
Glass solubility represents an important characteristic of
bioactive glasses ability to play a therapeutic role in vivo. A
review article by Hoppe et al.15
cites the relevance of inves-
tigating the effects of ionic dissolution products from bioac-
tive glasses. Si41
release increases bone mineral density
while inducing hydroxyapatite (HaP) precipitation and also
stimulates collagen I formation and osteoblastic differentia-
tion. Ca21
and PO32
4 release are also known to play roles in
bone metabolism such as encouraging osteoblast differentia-
tion and proliferation and extracellular matrix (ECM) syn-
thesis. P has been shown to stimulate expression of matrix
proteins critical to bone formation.15
Ion release profiles
determined high release rates for Si41
and Na1
which are
seen to reduce with the increase in TiO2 concentration
within the glass. The reduction in ion release for each TiO2
addition may be attributed to a number of factors. The BG-1
to BG-3 glasses have greater network disruption (from NMR
FIGURE 10. Surface of a) Control surface without adhered cells, b) BG-3 with adhered osteoblasts, and c) BG-5 also with adhered osteoblasts.
FIGURE 11. Higher resolution SEM image (31000) of osteoblast
adhered to surface of BG-5.
TABLE III. Growing Cells (G. Cells) Viability Compared to the
Cell Viability of Each Glass Composition at 1, 7, and 14 Days
Incubation in Aqueous Media (significant when p 0.05)
1 Day 7 Days 14 Days
G. Cells Control 0.941 1.000 1.000
BG1 1.000 1.000 1.000
BG3 1.000 1.000 1.000
BG5 1.000 0.203 1.000
1710 PLACEK ET AL. INVESTIGATING THE EFFECT OF TIO2

9. spectra) thereby reducing the BO species in addition to the
reduced Si41
concentration. Ti41
is a highly charged tetrava-
lent ion that may be utilizing Na1
as a charge compensator,
as Ca21
levels experience relatively minor changes with
respect to Ti41
concentration (Figure 7). The slight reduction
in Ca21
levels may be attributed to precipitation onto the
glass particles surfaces. The reduction in Si41
release is likely
due to the reduced Si41
concentration as a function of TiO2
substitution. Previous studies by Talos¸ et al. on solgel derived
xerogels investigated Ti41
effect on ion release and deter-
mined that the incorporation of TiO2 resulted in improving
solgel stability. Stability was achieved as hydration suscepti-
ble P-O-P bonds were partially replaced by hydration-
resistant P-O-Ti. Since Ti41
ions have a small ionic radius and
a large electric charge, they can be easily integrated into the
glassy network and form stronger P-O-Ti bonds compared to
P-O-P bonds.39
This may explain the lack in change with
respect to PO32
4 release. Additional studies on phosphorus-
based glasses also experienced similar characteristics as
described here when studying the effect on structure and sol-
ubility. This study found an increase in Tg (evident in BG-4
and BG-5) as the concentration of Ti41
in the glass
increased.21
This may support earlier finding by Talos¸ et al.
and in this study, where P-O-Ti are more resistant to hydra-
tion breakdown. There was also an associated reduction in
the degradation rate and ion release of the glasses which was
attributed to an increase in density. This resulted in a reduc-
tion in all ions released from the glass.21
Changes in pH were experienced with each material, in
particular as a function of incubation time (1, 7 and 14
days). Control experienced the highest pH at 12.2 after 14
days. BG-1, BG-2. and BG-3 presented pH ranging from 11.1-
11.6 after the same time period, while BG-4 and BG-5
presents pH of 10.3 and 10.4, respectively (Figure 8). This
trend closely follows changes in structure and ion release
where Control, which has the lowest Tg, greater distribution
of Q-species and highest ion release rate, experiences the
highest pH. The changes in pH are likely related to Si41
and
Na1
release as Ca21
and PO32
4 release levels are low in
comparison and do not experience significant changes with
respect to time. Cell viability studies (Figure 9) were con-
ducted on each glass and compared to a healthy growing
cell population with respect to maturation (1, 7, and 14
days). Statistical comparisons (Tables III and IV) were con-
ducted between glasses at each time period and as a func-
tion of maturation time. After 1 day incubation, cell viability
was reduced by 25% by the Control glass compared to the
growing cells, however, this reduction was not deemed sig-
nificant (p 5 0.941). There was also no significant differ-
ence for BG-1, BG-3, and BG-5 (p 5 1.000), even though BG-
5 increased cell viability by 21%. This initial reduction in
cell viability by the Control glass may be attributed to the
higher solubility, as evident with ion release (particularly
Si41
and Na1
release) and pH studies. Regarding the 7 day
samples, Control, BG-1 and BG-3 presented similar cell via-
bility levels which deviated from the growing cell popula-
tion by 65%. At 7 days, BG-5 presented cell viability 32%
greater than the growing cell population, however, this dif-
ference was also not significant compared to the growing
cells (p 5 0.203). At 14 days, all samples compared to the
growing cell population did not prove to be significant (p 5
1.000). Comparing 1 and 14 day samples for each glass,
there was no significant difference attributed to Control (p
5 0.215), BG-1 (p 5 0.941), BG-3 or BG-5 (p 5 1.000). Cell
culture studies determined that no cell toxicity was experi-
enced by any of the glass compositions evaluated. Cell adhe-
sion studies (Figures 10 and 11) supported viability studies
where the higher solubility of the Control glass after 24 h
resulted in reduced viability, in addition to no osteoblasts
adhering and proliferating on the glass surface. BG-3 and
BG-5 presented an increasing number of osteoblasts as the
solubility of the glass decreased. It has previously been
shown that ionic dissolution products from bioactive glasses
influence the behavior of cells in vivo by stimulating the
genes of the affected cells toward self-repair. It has been
shown that these ionic dissolution products from glasses
can stimulate osteogenesis and angiogenesis, processes that
make these materials highly desirable for bone tissue
repair.15
An additional characteristic which makes these
materials unique, is that there still exists a degree of uncer-
tainty as to which ionic dissolution products cause their
associated therapeutic responses, also, in what concentra-
tion, and what combination of released elements contribute
to these effects.
CONCLUSIONS
Incorporating Ti41
in a 42SiO2224Na2O-21CaO-13P2O5 wt
% bioactive glass resulted in modifying the structure of the
glasses, particularly at higher concentrations of 25 and 30
wt % TiO2. MAS-NMR data in particular, indicates a shift in
the role Ti41
plays within the glass at higher concentrations.
In addition, the solubility of the Ti41
containing glasses was
greatly reduced compared to the Control glass which also
resulted in reducing the solution pH as the Ti41
concentra-
tion increased. Additionally, cell viability was improved at
earlier time periods with the increase in Ti41
, which also
providing a more amenable surface for adhesion of MC-3T3
osteoblasts. Future work on these compositions will include
analyzing surface changes in simulated body fluid in order
to determine if the reduction in Si41
levels influences the
formation of an apatite layer on the glasses surface.
ACKNOWLEDGMENTS
The authors thank the following people for their assistance
with this project: C.M. Smith, C.M. Clark, M.W. Strohmayer, N.L.
Keenan, H.A. Liggett, H.F. Haddad, L.E. Jaeger, and L.R. Boutelle.
TABLE IV. Comparison of 1 and 14 Day Cell Viability Sam-
ples for Each Individual Glass Composition (Significant when
p 0.05)
1 vs. 14 Days
Control 0.215
BG1 0.941
BG3 1.000
BG5 1.000
ORIGINAL RESEARCH REPORT
JOURNAL OF BIOMEDICAL MATERIALS RESEARCH B: APPLIED BIOMATERIALS | NOV 2016 VOL 104B, ISSUE 8 1711

10. REFERENCES
1. Hench LL. The story of bioglass. J Mater Sci: Mater Med 2006;17:
967–978.
2. Jones JR. Review of bioactive glass: From Hench to hybrids. Acta
Biomater 2013;9:4457–4486.
3. Chen QZ, Thompson ID, Boccaccini AR. 45S5 BioglassVR -derived
glass-ceramic scaffolds for bone tissue engineering. Biomaterials
2006;27:2414–2425.
4. Vargas GE, Mesones RV, Bretcanu O, Lopez JMP, Boccaccini AR,
Gorustovich A. Biocompatibility and bone mineralization potential
of 45S5 BioglassVR -derived glass-ceramic scaffolds in chick
embryos. Acta Biomater 2009;5:374–380.
5. Haimi S, Gorianc G, Moimas L, Lindroos B, Huhtala H, Raty S,
Kuokkanen H, Sandor GK, Schmid C, Miettinen S, Suuronen R.
Characterization of zinc-releasing three-dimensional bioactive
glass scaffolds and their effect on human adipose stem cell prolif-
eration and osteogenic differentiation. Acta Biomater 2009;5:
3122–3131.
6. Wren AW, Cummins NM, Laffir FR, Hudson SP, Towler MR. The
bioactivity and ion release of titanium-containing glass polyalke-
noate cements for medical applications. J Mater Sci: Mater Med
2011;22:19–28.
7. Wren AW, Cummins NM, Towler MR. Comparison of antibacterial
properties of commercial bone cements and fillers with a zinc-
based glass polyalkenoate cement. J Mater Sci 2010;45:5244–
5251.
8. Wren AW, Boyd D, Towler MR. The processing, mechanical prop-
erties and bioactivity of strontium based glass polyalkenoate
cements. J Mater Sci: Mater Med 2005;19:1737–1743.
9. Anderson JH, Goldberg JA, Bessent RG, Kerr DJ, McKillop JH,
Stewart I, Cooke TG, McArdle CS. Glass yttrium-90 microspheres
for patients with colorectal liver metastases. Radiother Oncol
1992;25:137–139.
10. da Costa Guimaraes C, Moralles M, Roberto Martinelli J. Monte
Carlo simulation of liver cancer treatment with 166Ho-loaded
glass microspheres. Radiat Phys Chem 2014;95:185–187.
11. Bortot MB, Prastalo S, Prado M. Production and characterization
of glass microspheres for hepatic cancer treatment. Proc Mater
Sci 2012;1:351–358.
12. Silver IA, Deas J, Erecinska M. Interactions of bioactive glasses
with osteoblasts in vitro: Effects of 45S5 Bioglasses, and 58S and
77S bioactive glasses on metabolism, intracellular ion concentra-
tions and cell viability. Biomaterials 2001;22:175–185.
13. Kokubo T, Kim H-M, Kawashita M. Novel bioactive materials with
different mechanical properties. Biomaterials 2003;24:2161–2175.
14. Kokubo T, Takadama H. How useful is SBF in predicting in vivo
bone bioactivity. Biomaterials 2006;27:2907–2915.
15. Hoppe A, Guldal NS, Boccaccini AR. A review of the biological
response to ionic dissolution products from bioactive glasses and
glass-ceramics. Biomaterials 2011;32:2757–2774.
16. Saffarian Tousi N, Velten MF, Bishop TJ, Leong KK, Barkhordar
NS, Marshall GW, Loomer PM, Aswath PB, Varanasi VG. Combi-
natorial effect of Si41, Ca21 and Mg21 released from bioactive
glasses on osteoblast osteocalcin expression and biomineraliza-
tion. Mater Sci Eng 2013;33:2757–2765.
17. Hench LL. Genetic design of bioactive glass. J Eur Ceram Soc
2009;29:1257–1265.
18. Takadama H, Kim H-M, Kokubo T, Nakamura T. XPS study of the
process of apatite formation on bioactive Ti-6Al-4V alloy in simu-
lated body fluid. Sci Technol Adv Mater 2001;2:389–396.
19. Gonzalez JEG, Mirza-Rosca JC. Study of the corrosion behavior of
titanium and some of its alloys for biomedical and dental implant
applications. J Electroanal Chem 1999;471:109–115.
20. Lausmaa J. Surface spectroscopic characterization of titanium
implant materials. J Electron Specrosc Relat Phenom 1996;81:
343–361.
21. Lakhkar NJ, Lee I-H, Kim H-W, Salih V, Wall B, Knowles JC. Bone
formation controlled by biologically relevant inorganic ions: Role
and controlled delivery from phosphate-based glasses. Adv Drug
Deliv Rev 2013;65:405–420.
22. Piscanec S, Ciacchi LC, Vesselli E, Comelli G, Sbaizero O, Meriani
S, De Vita A. Bioactivity of TiN-coated titanium implants. Acta
Mater 2004;52:1237–1245.
23. Iwamoto N, Tsunawaki Y, Masao F, Hatfori T. Raman spectra of
K2O-SiO2 and K2O-SiO2-TiO2 glasses. J Non-Cryst Solids 1975;
18:303–306.
24. Kusaeiraki K. Infrared and Raman spectra of vitreous silica and
sodium silicates containing titanium. J Non-Cryst Solids 1987;
9596:411–418.
25. Satyanarayana T, Kityk IV, Ozga K, Piasecki M, Bragiel P, Brik MG,
Ravi Kumar V, Reshak AH, Veeraiah N. Role of titanium valence
states in optical and electronic features of PbO-Sb2O3-B2O3:TiO2
glass alloys. J Alloys Compd 2009;482:283–297.
26. Wang JY, Wicklund BH, Gustilo RB, Tsukayama DT. Titanium,
chromium and cobalt ions modulate the release of bone-
associated cytokines by human monocytes/macrophages in vitro.
Biomaterials 1996;17:2233–2240
27. Liu HC, Chang WH, Lin FH, Lu KH, Tsuang YH, Sun JS. Cytokine
and prostaglandin E2 release from leukocytes in response to
metal ions derived from different prosthetic materials: An in vitro
study. Artif Organs 1999;23:1099–1106.
28. Chan EP, Mhawi A, Clode P, Saunders M, Filgueira L. Effects of
titanium(iv) ions on human monocyte-derived dendritic cells. Met-
allomics 2009;1:166–174.
29. Sun ZL, Wataha JC, Hanks CT. Effects of metal ions on
osteoblast-like cell metabolism and differentiation. J Biomed
Mater Res 1997;34:29–37.
30. Liao HH, Wurtz T, Li JG. Influence of titanium ion on mineral for-
mation and properties of osteoid nodules in rat calvaria cultures.
J Biomed Mater Res 1999;47:220–227.
31. Mine Y, Makihira S, Nikawa H, Murata H, Hosokawa R, Hiyama A,
Mimura S. Impact of titanium ions on osteoblast-, osteoclast- and
gingival epithelial-like cells. J Prosthodont Res 2010;54:1–6.
32. Wren AW, Laffir FR, Kidari A, Towler MR. The structural role of
titanium in Ca-Sr-Zn-Si/Ti glasses for medical applications. J Non-
Cryst Solids 2011;357:1021–1026.
33. Serra J, Gonzalez P, Liste S, S. Chiussi, Leon B, Perez-amor M,
Ylanen HO, Hupa M. Influence of the non-bridging oxygen groups
on the bioactivity of silicate glasses. J Mater Sci: Mater Med
2002;13:1221–1225.
34. Calas G, Cormier L, Galoisy L, Jollivet P. Structure-property rela-
tionships in multicomponent oxide glasses. Comp Rend Chim
2002;5:831–843.
35. Wren AW, Coughlan A, Smith CM, Hudson SP, Laffir FR, Towler
MR. Investigating the solubility and cytocompatibility of CaO–
Na2O–SiO2/TiO2 bioactive glasses. J Biomed Mater Res: Part A
2014;DOI: 10.1002/jbm.a.35223.
36. Aguiar H, Serra J, Gonzalez P, Leon B. Structural study of sol–gel
silicate glasses by IR and Raman spectroscopies. J Non-Cryst Sol-
ids 2009;355:475–480.
37. Galliano PG, Porto JM, Spezl L, Varetti EL, Sobrados I, Sanz J.
Analysis by nuclear magnetic resonance and Raman spectroscop-
ies of the structure of bioactive alkaline-earth silicophosphate
glasses. Mater Res Bull 1994;29:1297–1306.
38. Hayakawa S, Osaka A, Nishioka H, Matsumoto S, Minura Y. Struc-
ture of lead oxyfluorosilicate glasses: X-ray photoelectron spec-
troscopy and nuclear magnetic resonance spectroscopy and
molecular dynamics simulation. J Non-Cryst Solids 2000;272:103–
118.
39. Talos¸ F, Senila M, Frentiu T, Simon S. Effect of titanium ions on
the ion release rate and uptake at the interface of silica based
xerogels with simulated body fluid. Corros Sci 2013;72:41–46.
1712 PLACEK ET AL. INVESTIGATING THE EFFECT OF TIO2