Chitosan adsorption on hydroxyapatite and its role in preventing acid erosion

Hyun-Su Lee
a
, Shannon Tsai
a
, Chin-Chen Kuo
a
, Alice W. Bassani
a
, Brian Pepe-Mooney
a
, Davide Miksa
b
,
James Masters
b
, Richard Sullivan
b
, Russell J. Composto
a,
a
Department of Materials Science and Engineering, University of Pennsylvania, Philadelphia, PA 19104, United States
b
Colgate-Palmolive Company, Piscataway, NJ 08855, United States
a r t i c l e i n f o
Article history:
Received 21 April 2012
Accepted 25 June 2012
Available online 7 July 2012
Keywords:
Chitosan
Hydroxyapatite
HA
Acid erosion
In situ quartz-crystal microbalance with
dissipation
QCM-D
Atomic force microscopy
AFM
Polymer adsorption
a b s t r a c t
Polymer adsorption onto an articial saliva (AS) layer is investigated using quartz-crystal microbalance
with dissipation (QCM-D) and chitosan as the model polymer. QCM-D is utilized in an innovative manner
to monitor in situ adsorption of chitosan (CH) onto a hydroxyapatite (HA) coated crystal and to examine
the ability of the adsorbed layer to protect the HA upon sequential exposure to acidic solutions. After
deposition of a thin AS layer (16 nm), the total thickness on the HA substrate increases to 37 nm upon
exposure to CH at pH 5.5 for 10 min. Correspondingly, the surface charge changes from negative (i.e.,
AS) to positive, consistent with the adsorption the polycationic CH onto or into the AS layer. Upon expo-
sure to an oxidizing agent, the chitosan cross-links and collapses as noted by a decrease in thickness to
10 nm and an increase in the shear modulus by an order of magnitude. Atomic force microscopy (AFM) is
used to determine the surface morphology and RMS roughness of the coated and HA surfaces after citric
acid challenges. Both physisorbed and cross-linked chitosan are demonstrated to limit and prevent the
erosion of HA, respectively.
2012 Elsevier Inc. All rights reserved.
1. Introduction
Over the last decade, dental enamel erosion in all age groups
has increased due to a variety of dietary factors [1]. One of the
main causes has been attributed to an increase in the consumption
of beverages containing citric acid, such as citrus fruit juices and
carbonated soft drinks [2]. Permanent dental enamel is primarily
composed of mineralized carbonated hydroxyapatite (Ca
10x
Na
x
(PO
4
)
6y
(CO
3
)
2
(OH)
2u
F
u
), and can be modeled using calcium
hydroxyapatite (HA), Ca
10
(PO
4
)
6
(OH)
2
, due to their similar compo-
sition and structure. As with enamel, direct exposure of HA to acids
results in demineralization [3,4]. In vivo, saliva provides limited
protection against this erosion by forming a pellicle layer that coats
the HA surface [5,6]. In order to provide a better means of protec-
tion, the present study examines the use of a polymer coating to
further deter the effects of acid erosion of HA. Of particular interest
is the ability of polymers to modify the native properties of the
host surface by changing the wettability [7], surface topography
[8], or chemical reactivity [9]. One goal of this study is to show that
model experimental systems can allow for the elucidation of poly-
mer interactions with surfaces [10,11].
Normally, HA coated with a pellicle layer undergoes demineral-
ization and re-mineralization reactions depending on the ion prod-
uct. In the idealized chemical reaction, as the pH decreases,
positive hydrogen ions from the acid bind with the negative phos-
phate and hydroxyl ions from the HA (enamel). As a result, the io-
nic solution in the pellicle layer becomes unsaturated resulting in a
shift that favors demineralization, which leads to the loss of cal-
cium and phosphate ions from the crystal until a solubility equilib-
rium is reached (Fig. 1). Dissolution, however, only occurs after
diffusion of the acid through the pellicle and the protein-lipid coat-
ing. The details of acid erosion in enamel and dentin are described
by Featherstone and Magalhes et al. [12,13].
To prevent dental erosion, several strategies have been utilized
including increasing saliva or plaque calcium by calcium treat-
ments to counteract inherent deciencies in the crystal structure
[13], promoting re-mineralization and strong teeth by uoride
exposure [14], and using laser irradiation to improve the resistance
of enamel to acid [15]. Recently, however, many products have
been developed that utilize polymers as a protective coating
against acid exposure. Polymer coatings serve as an efcient meth-
od to protect surfaces from erosion because they can be biocom-
patible and have been known to be effective agents against
0021-9797/$ - see front matter 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcis.2012.06.074

Corresponding author. Address: Department of Materials Science and Engi-

neering, University of Pennsylvania, Laboratory for Research on the Structure of
Matter, 3231 Walnut Street, Philadelphia, PA 19104, United States.
E-mail addresses: Hyun-Su.Lee@uphs.upenn.edu (H.-S. Lee), shantsai@seas.
upenn.edu (S. Tsai), janson925@gmail.com (C.-C. Kuo), awbassani@gmail.com
(A.W. Bassani), bpepemoo@gmail.com (B. Pepe-Mooney), ume_2000@yahoo.com
(D. Miksa), Jim_Masters@colpal.com (J. Masters), Richard_Sullivan@colpal.com
(R. Sullivan), composto@seas.upenn.edu (R.J. Composto).
Journal of Colloid and Interface Science 385 (2012) 235243
Contents lists available at SciVerse ScienceDirect
Journal of Colloid and Interface Science
www. el sevi er . com/ l ocat e/ j ci s
corrosive factors. To date, polymer coated HA surfaces have been
shown to be very successful in combating acid erosion [1618].
In the present study, the polycationic biopolymer, chitosan, was
chosen due to its adhesion properties to both hard and soft sur-
faces and potential dental applications [19]. While its unique
mucoadhesive and therapeutic properties is evident in the litera-
ture [9,10,20], to our knowledge, the benets of a chitosan modi-
ed-HA surface in preventing acid erosion has not yet been
examined. Chemically, chitosan is produced by the controlled N-
deacetylation of chitin, a linear polysaccharide sourced primarily
from the exoskeleton of arthropods such as marine crustaceans
[21]. The resulting polycation is soluble in aqueous solutions of
small organic acids [2225], such as acetic and lactic acids, and
can be cross-linked in the presence of polyvalent anions such as
phosphates [26].
A variety of surface characterization methods have been used to
study the adsorption and layer growth of polymers on surfaces.
Several techniques are capable of directly observing lm formation
at the interface between surface and solution. In situ studies of the
adsorption and formation of polymers onto surfaces have been re-
ported to date and described in further detail in a review by Ober
et al. [27]. These methods include using atomic force microscopy
[2830], internal reection spectroscopy [31,32], surface plasmon
resonance [33], spectroscopic ellipsometry [34], and quartz crystal
microbalance with dissipation monitoring (QCM-D) [3552]. One
signicant advantage of QCM-D is that it allows for in situ mea-
surements of both the polymer lm modulus and viscoelasticity.
The study of the interfacial properties of polymer/surface systems
has been greatly aided by the introduction of the quartz microbal-
ance technique. QCM has been successfully employed in the study
of layer-by-layer polymer deposition as well as in the study of
deposition processes of inorganic materials [53,54]. This technique
is sensitive to changes in both mass and viscoelasticity without
requiring the addition of labels to the adsorbing species. For these
reasons, QCM was chosen as the primary end-point measurement
for the study herein described as a means of real-time monitoring
of chitosan physisorption onto HA, chitosan cross-linking, and HA
erosion.
Although QCM-D has been used to examine polymer adsorption
onto surfaces, this study applies QCM to monitor the entire
sequence of steps designed to simulate in situ dental erosion. This
includes the monitoring of AS adsorption, chitosan adsorption on
the AS with and without cross-linking, and nally acid erosion
tests against HA. For the present study, QCM-D was used to
observe, in-situ, sequential changes in absorption involving a thin
layer of AS (10 nm) adsorbed onto an HA crystal followed by a
second step where chitostan is incorporated into the AS. The mass
sensitivity of QCM-D is 1.77 ng/cm
2
(0.1 Hz) [48], which makes it a
useful tool for these measurements. For the purposes of our stud-
ies, rather than using a thicker layer of AS (pellicle), the thinner
layer of AS was required in order to accurately perform the
QCM-D measurements due to the sensitivity of the procedure.
In the present study, we show that chitosan physically adsorbs
to the AS coated HA surface. Using QCM-D, we owed a 10 wt.% AS
solution over an HA coated QCM-crystal, until steady state adsorp-
tion was observed. Next, the excess was rinsed and a 0.5 wt.% solu-
tion of chitosan was owed over the sample to obtain a stable
polymer coating. We then show that by exposing the adsorbed
chitosan to sodium periodate (NaIO
4
), an oxidizing agent, the ini-
tially physisorbed chitosan layer (thickness 37 nm) collapses to
form a stiffer, more solid-like lm (thickness 10 nm). These results
were veried using AFM.
We also analyzed the ability of both the brushy and collapsed
chitosan surface to prevent acid erosion of HA. Using QCM-D, the
chitosan surfaces were subjected to 10 cycles of a 5 min acid expo-
sure followed by a 5 min rinse. The brushy chitosan surface
showed some deterioration with each cycle, while the collapsed
chitosan showed no signicant change in thickness and viscoelas-
ticity over the 10 cycles. These results were conrmed through fur-
ther examination using AFM. Using a plasma etching method to
remove the adsorbed layers [55], we were able to directly evaluate
the surface morphology and roughness of the HA surface after acid
challenges. Following the acid challenges and removal of all ad-
sorbed species, the HA surface morphology of the brushy chitosan
showed some changes in morphology, suggesting that some ero-
sion had occurred. However, when the surface morphology of the
cross-linked chitosan was examined, no signicant changes in sur-
face morphology were observed. We found that while physically
adsorbed chitosan provided some protection to acid challenges,
the cross-linked chitosan served as a better coating in the preven-
tion of acid erosion of HA. These results were in agreement with
QCM-D results. As such, it was also determined that QCM-D is a no-
vel and benecial technique for the in-situ monitoring of HA
erosion.
2. Experimental section
2.1. Materials
QCM sensor crystals, an AT-cut piezoelectric quartz crystal
(14 mm in diameter and 0.3 mm thickness) coated with uniform
nanocrystalline hydroxyl apatite (Ca
10
(PO
4
)
6
(OH)
2
, effective HA
coating: 10 nm) by Promimic AB, were purchased from Biolin
Fig. 1. Chemical reaction conditions corresponding to re-mineralization of enamel (left), equilibrium between enamel production and degradation, and demineralization of
enamel due to acidic exposure(right). Bottom: The conditions where ion product (IP = [Ca
2+
]
10
[PO
3
4
]
6
[OH

]
2
) is greater than, equal to or less than the solubility product
constant of HA, K
0
sp
. The pellicle is a protein rich layer that forms on enamel due to salivary action.
236 H.-S. Lee et al. / Journal of Colloid and Interface Science 385 (2012) 235243
Scientic, Inc. USA. The mechanical strength and stability are also
found to be similar (Kjellin P, 2006). N-type, (100) oriented silicon
wafers (CZ silicon: dopant, Ph: 2030 X resistivity) were pur-
chased from Silicon Quest International, USA. Chitosan Chitoclear

Cg-10 (Primex ehf. Iceland, M

w
= 60 kDa and degree of deacetyla-
tion: 87%) was kindly donated by Colgate-Palmolive Co. (Piscata-
way NJ). Sodium periodate (P99.8%) and citric acid (P99.5%)
were purchased from the Sigma-Aldrich Chemical Co. USA. All re-
agents for the articial saliva solution (mucin from porcine stom-
ach (type II), ammonium chloride (NH
4
Cl, P99.5%), calcium
chloride dihydrate (CaCl
2
2H
2
O), magnesium chloride hexahydrate
(MgCl
2
6H
2
O), potassium chloride (KCl), potassium phosphate
monobasic (KH
2
PO
4
), potassium thiocyanate (KCNS), sodium bicar-
bonate (NaHCO
3
), sodium phosphate dibasic (Na
2
HPO
4
), sodium
citrate dihydrate (Na
3
C
6
H
5
O
7
2H
2
O), albumin (BSA), urea, glycine,
sodium azide (NaN
3
)) were also purchased from the Sigma-Aldrich
Chemical Co. USA. Acetic acid, glacial, was purchased from Fisher
scientic. USA.
2.2. In-situ QCM-D measurement
An E4 Quartz Crystal Microbalance (QCM) instrument (Q-Sense
Inc., Gothenburg, Sweden) capable of dissipation monitoring was
used to quantify both polymer adsorption on HA surfaces as well
as acid erosion of the same. QCM-D is a powerful tool for measur-
ing the mass deposited onto or released from a surface in contact
with a liquid. The QCM-D is based on the change in resonant fre-
quency of a vibrating quartz crystal sensor, a piezoelectric mate-
rial, according to mass changes of the sensor. The QCM-D
instrument monitors real-time changes in the frequency of vibra-
tional modes as well as changes in the vibrational energy dissipa-
tion. The Sauerbrey equation [43,44] relates the change in the
resonant frequency to the change in mass of the quartz crystal,
Dm C
Df
n
n
where C is the mass sensitivity constant (C = 17.7 ng cm
2
Hz
1
for
an AT-cut, 5 MHz crystal) and n is the vibrational mode number
(n = 1, 3, 5, . . .). The dissipation change, DD
n
, indicates physical
characteristics of the deposited layer such as viscosity and elastic-
ity. If DD
n
is less than 2.0 10
6
and the plots of Df
n
/n over time
for several modes superimpose, the adsorbed layer behaves like
an elastic solid. Thus, the physical properties (mass and thickness)
of the elastic layer can be calculated using the Sauerbrey equation
[4347]. On the contrary, if DD
n
is greater than 2.0 10
6
and
the plots of Df
n
/n over time for several modes do not superimpose,
the layer behaves like a viscoelastic liquid. In addition, DD
n
indi-
cates the conformation changes of a deposited layer as a result of
physical or chemical treatments. An increase of DD
n
indicates that
the deposited layer has become more viscous whereas a decrease
of DD
n
indicates that the deposited layer has become more elastic.
Thus the change in DD
n
reects the viscoelastic behavior during
physical/chemical treatment. In this case, the physical properties
(thickness, shear modulus, and viscosity) of the viscoelastic layer
can be estimated by tting the QCM-D experimental data (Df
n
/n
(n = 1, 3, 5, . . .) and DD
n
) with a Voigt-based viscoelastic model
[39,4852]. Hydroxyapatite-coated QCM sensor crystals were
cleaned by UV-Ozone treatment for 20 min., immersed in 95% eth-
anol for 30 min. at room temperature, rinsed with ultrapure water
(Millipore Direct-Q, 18 mXcm resistivity), dried with N
2
(g), and
then re-exposed to UV-Ozone for 20 min. to remove organic and
biological impurities. The pH and ionic strength (IS) of the solutions
were measured with a dual pH/conductivity meter (Denver Instru.
Co. USA) and a conductivity meter (Mettler-Toledo, Switzerland),
respectively. For QCM-D studies, we used an articial saliva solu-
tion, which contains only two proteins (mucin and albumin)
[56,57], because the composition of this layer is less complex than
Fig. 2. Schematic drawing showing the oxidation and cross-linking of chitosan using the oxidizing reagent NaIO
4
. Oxidation produces aldehyde groups which react with the
primary amines to form intra and inter chain cross links in chitosan.
H.-S. Lee et al. / Journal of Colloid and Interface Science 385 (2012) 235243 237
that of the many component salivary lm [58]. First, the porcine
mucin and the salts solutions were made separately. The porcine
mucin was dissolved over time using mild heating in a ventilated
hood. The two solutions were then mixed to make articial saliva
solution (2.5% w/w mucin from porcine stomach (type II), 0.047%
w/w ammonium chloride (NH
4
Cl, P99.5%), 0.042% w/w calcium
chloride dihydrate (CaCl
2
2H
2
O), 0.007% w/w magnesium chloride
hexahydrate (MgCl
2
6H
2
O), 0.231% w/w potassium chloride (KCl),
0.071% w/w potassium phosphate monobasic (KH
2
PO
4
), 0.044% w/w
potassium thiocyanate (KCNS), 0.107% w/w sodium bicarbonate
(NaHCO
3
), 0.075% w/w sodium phosphate dibasic (Na
2
HPO
4
),
0.003% w/w sodium citrate dihydrate (Na
3
C
6
H
5
O
7
2H
2
O), 0.005%
w/w albumin (BSA), 0.035% w/w urea, 0.090% w/w glycine, 0.059%
w/w sodium azide (NaN
3
)) as directed in the literature [59,60].
The mixed solution was autoclaved to prevent bacterial contamina-
tion. In this study, 10% v/v articial saliva solution (pH 6.41, ionic
strength = 539 ls/cm) was prepared by using the mixed articial
saliva solution and DI water (pH = 5.68, IS = 1.74 ls/cm). For
QCM-D experiments, surfaces were exposed to 0.5 wt.% chitosan
in 0.1% acetic acid aqueous solution (pH = 5.5, IS = 622 ls/cm), as
well as 1 wt.% citric acid solution (pH = 2.8, IS = 841 ls/cm). For a
cross-linked chitosan/AS layer, 2% w/w sodium periodate solution
(pH 5.16, IS = 202 ls/cm) was prepared. All solutions were degassed
by sonication. The liquid medium was circulated by peristaltic
pump at a rate of 0.1 mL/min through a ow cell with the sensor
crystal. The temperature of the system was maintained at 37 C.
To prepare a cross-linked chitosan/AS layer on the HA surface,
the physisorbed chitosan on the AS layer was exposed to 2 wt.% so-
dium periodate solution for 15 min at a ow rate of 0.1 mL/min.
The periodate oxidation of vicinal diols [61] (glucose and galactose
units of mucins, heavily glycosylated proteins, in AS layer) and vic-
inal amino alcohols [62,63] (GlcN residues in chitosan) leads to the
formation of dialdehyde functional groups. These aldehyde func-
tional groups interact with the residual primary amine groups of
GlcN and therefore result in the crosslinking of the chitosan ad-
sorbed on the HA layer (Fig. 2). QCM-D experiments (data not
shown) show that exposing an AS layer on HA to NaIO
4
does not
alter the thickness or viscoelasticity of the AS layer. This control
study supports the initial interpretation that the collapse of the
chitosan/AS layer upon exposure to the oxidizing agent is due to
the cross-linking of chitosan. In order to evaluate the ability of
the brushy (i.e., physisorbed chitosan/AS layer) and collapsed
chitosan (i.e., cross-linked chitosan/AS layer) surfaces to prevent
acid erosion of HA using QCM-D, the chitosan surfaces were sub-
jected to 10 cycles of a 5 min acid exposure (citric acid solution
(pH = 2.8)) followed by a 5 min rinse.
2.3. AFM measurements
Surface topography and roughness were determined by atomic
force microscopy (Digital Instruments, Santa Barbara, CA: Dimen-
sion 3000 AFM). Tapping mode with a single crystal Si tip with a
spring constant of 48 N/m, a radius of curvature of about 10 nm,
and a resonance frequency of approximately 190 kHz were used.
AFM images were taken over a scan size of 1 1 lm
2
. Images were
analyzed using Picoview 1.6 software (Agilent Technologies). The
RMS roughness values were determined from six separate 1 lm
2
images for each substrate type.
To characterize the HA surface after adsorption of chitosan/AS, a
method was developed to remove the organic layer by exposure to
oxygen plasma. A plasma cleaner/Sterilizer (PDC-001, Harrick Sci-
entic Corp. USA) was used to remove the organic lm without
damaging the underlying HA surface. A vacuum pump with a min-
imum pump speed of 1.4 m
3
/h (0.83 ft
3
/min) and an ultimate total
pressure of 200 m Torr (0.27 mbar) with pure oxygen gas were
used. The surface was cleaned using 29.6 W of power input for
15 min. As a control, AFM scans were taken from the same HA sur-
face before and after oxygen plasma exposure. The morphology
was similar and the RMS roughness was 3% of the original value.
This study showed that the HA surface was stable and did not dam-
age during exposure to oxygen plasma. Second, we prepared a
physisorbed chitosan layer on the HA surface to demonstrate that
the oxygen plasma treatment (15 min) can remove the organic lm
without damaging the HA surface. The morphology of the HA sur-
face after oxygen plasma treatment was found to be similar to the
original HA surface with an RMS within 5%. This shows that the
oxygen plasma removes the organic lm from the surface without
roughening or damaging the HA surface. Thus, after acid exposure
of the chitosan/AS coated HA surface, oxygen plasma treatment
can be used to reveal the HA surface, and any changes to the sur-
face features can be attributed to acid erosion.
2.4. Surface zeta potential measurement
The zeta-potentials of the AS layer, CH layer, and CH/AS layers
on silicon oxide substrates were measured using a DelsaNano-C
(Beckmann Coulter) instrument [50]. DI water (Millipore Direct-
Q, 18 mXcm resistivity, pH 5.5) was used in each measurement.
3. Results and discussion
Two model surfaces (physisorbed chitosan/AS layer (C) and
cross-linked chitosan/AS layer (D) in Fig. 3) were prepared in order
to investigate the adsorption of chitosan onto a hydroxyapitate
(HA) surface and to examine the ability of these coatings to prevent
acid erosion of HA. As shown in Fig. 3, rst, an articial saliva layer
was adsorbed onto the HA lm on the QCM sensor. Then this sur-
face was exposed to a solution of chitosan, which was subse-
quently exposed to NaIO
4
and cross-linked to create a tightly
packed layer on the HA. One objective of this study was to deter-
mine if the collapsed chitosan (D) was more effective at protecting
the HA surface from acid induced erosion relative to physisorbed
chitosan (C). Thus the nal stage of research was to characterize
the surface of the HA surface after acid exposure. This characteriza-
tion was performed with the use of QCM-D and AFM.
Fig. 3. Surface treatment illustrating the physical adsorption of chitosan onto an articial saliva (AS) layer deposited on hydroxyapatite (HA) coated QCM sensor. Acid
challenges were performed on both chitosan/AS (C) and cross-linked chitosan/AS surfaces (D). QCM-D was used to determine the mass adsorbed and corresponding
viscoelastic properties during each step. AFM was used to characterize the surface of the HA after removing chitosan/AS and cross-linked chitosan/AS using oxygen plasma.
238 H.-S. Lee et al. / Journal of Colloid and Interface Science 385 (2012) 235243
3.1. Monitoring the physisorption of chitosan and the cross-linking of
chitosan
The physisorption of chitosan on HA surfaces, pretreated with
AS, was studied using in-situ QCM-D. After using DI water to estab-
lish a baseline, the HA surface of the QCM sensor was exposed to a
solution of articial saliva (AS) (arrow 1 in Fig. 4). For the three
vibrational modes (n = 3, 5, 7), Df
n
/n decreases rapidly and thereaf-
ter each mode becomes distinct; correspondingly DD
n
increases to
4 10
6
. These results are consistent with an adsorbed articial
saliva layer that behaves viscoelastically. After rinsing the articial
saliva layer with DI water (arrow 2 in Fig. 4), Df
n
/n (n = 3, 5, 7)
slightly increases and then remains constant as the modes super-
impose, while DD
n
decreases to 2 10
6
. This behavior suggests
that the ow of DI water removes loosely bound species, resulting
in a nal articial saliva layer that is thinner and more elastic than
the initial physisorbed layer. At this stage of the process, the HA
surface is coated with a stable layer of articial saliva.
For chitosan exposure on the articial saliva layer, similar to
Dedinate et.al researches [64], an acidic chitosan solution
(pH = 5.5, IS = 622 ls/cm) was used because chitosan is insoluble
in high pH (above 6.5). After the articial saliva layer is exposed
to a chitosan solution (arrow 3 in Fig. 4), Df
n
/n (n = 3, 5, 7) de-
creases and the modes show their largest separation; correspond-
ingly, DD
n
increases to its largest value, greater than 10 10
6
.
These results indicate that chitosan has physically adsorbed on,
or becomes incorporated into, the articial saliva layer and that
the organic layer on the HA surface is viscoelastic (i.e., behaves
as both elastic solid and viscous liquid). To conrm that CH ab-
sorbed on (into) the AS layer, the surface charge of the AS, CH,
and CH/AS layers was determined by measuring their zeta poten-
tials. The zeta potentials of the AS and CH layers are 10 mV and
+25 mV at pH 5.5, respectively. This result is consistent with values
reported in the literature [50,65,66]. Interestingly the zeta poten-
tial of physisorbed CH/AS layer is +9 mV, a value between the AS
and CH layers. The change in sign and intermediate value is a good
evidence that positively charged chitosan adsorbs onto or into the
negative charged AS layer. Previous studies by Dedinaite et al.
[64,67] demonstrated that the positively charged chitosan associ-
ates with a pre-adsorbed mucin layer to form mucin-chitosan
complexes. In our study, because the mucin concentration is
500x that of BSA, the AS layer, as evidenced by its negative charge,
is mainly mucin. Therefore the lm on the QCM crystal is likely a
mixture of chitosan with the AS.
To prepare a cross-linked chitosan/AS layer, the physisorbed
CH/AS layer is exposed to a NaIO
4
solution (arrow 4 in Fig. 4). As
a result, the Df
n
/n (n = 3, 5, 7) curves superimpose again, after an
initial transient, as shown between 50 and 55 min. Note that the
frequency change before and after exposing the chitosan/AS layer
to NaIO
4
is relatively small whereas the decrease in DD
n
is rela-
tively large, 3 10
6
, immediately after introducing the oxidiz-
ing agent. Thus, although the mass change on the HA crystal is
small, the chitosan/AS layer behaves like an elastic solid because
of the collapse of the organic layer due to the cross-linking of chito-
san. After rinsing the cross-linked chitosan/AS layer with DI water
(arrow 5 in Fig. 4), Df
n
/n (n = 3, 5, 7) increases slightly and then
reaches a constant value of about 30 Hz for all modes (i.e, curves
superimpose), whereas DD
n
decreases to its lowest value, less than
1 10
6
. Thus, rinsing with DI water removes weakly bound spe-
cies. The remaining layer contains cross-linked chitosan integrated
with the AS, forming a stable, thin and elastic lm on HA.
In separate experiments, physisorbed chitosan layers were pre-
pared by replacing the oxidation step (arrow 4 in Fig. 4) with
15 min. of rinsing using DI water (not shown) to remove weakly
bound species. In this case Df
n
/n (n = 3, 5, 7) increases to 25 Hz,
the modes superimpose, and DD
n
decreases to 2 10
6
. Thus,
Fig. 4. QCM-D results showing the behavior of Df
n
/n (n = 3, 5, 7) and DD
n
for times
corresponding to (1) formation of articial saliva layer, (2) DI water rinsing of the
articial saliva layer, (3) physical adsorption of chitosan into articial saliva layer,
(4) exposure of chitosan/articial saliva to an oxidizing agent, NaIO
4
, that cross-
links chitosan, and (5) DI water rinsing of the cross-linked chitosan/articial saliva
layer. A decrease in frequency corresponds to an increase in surface thickness.
Fig. 5. (a) QCM-D data showing Df
n
/n (n = 3, 5, 7) and DD
n
as a function of exposure
time. The HA surface is modied by an articial saliva layer, then exposed to
chitosan which is then cross-linked. The cross-linked chitosan/AS layer time is then
exposed to citric acid followed by rinsing with water. The dashed lines represent
the ts to the data and are used to determine the lm thickness in (b). (b) The
thickness of the articial saliva (16 nm), articial saliva after exposure to chitosan
(37 nm) and chitosan/AS layer after cross-linking (10 nm) are shown. Arrows 6 and
7 represent times corresponding to the initial exposure of the cross-linked chitosan
to citric acid and pure water, respectively. The thickness of cross-linked chitosan
remains constant after ten acid challenges.
H.-S. Lee et al. / Journal of Colloid and Interface Science 385 (2012) 235243 239
the exposure of the physisorbed chitosan/AS layer to DI water re-
moves loosely bound components resulting in a stable, elastic,
thin-lm. This lm is thicker and more viscous than that of the
cross-linked chitosan/AS layer.
3.2. Effect of acid exposure to chitosan/AS on HA using QCM-D
The stability of the physisorbed chitosan/AS and the cross-
linked chitosan/AS layers on HA against acid challenges was inves-
tigated using in-situ QCM-D. Fig. 5 shows how Df
n
/n (n = 3, 5, 7)
and DD
n
vary during the formation of AS layer, cross-linking of
chitosan/AS layer, and exposure of the chitosan/AS layer to 10 acid
challenges. The initial treatment of the HA surface was described
when discussing Fig. 4 and will be again analyzed in the context
of Fig. 5. Each acid challenge consists of exposing CH/AS to citric
acid (pH 2.8) for 5 min followed by a DI water rinse for 5 min. After
the cross-linked chitosan/AS layer is exposed to citric acid (arrow
6, Fig. 5a), Df
n
/n (n = 3, 5, 7) decreases to about 35 Hz for all
modes and the curves slightly separate, whereas DD
n
increases
to 3 10
6
. After rinsing with DI water (arrow 7 in Fig. 5 (a)),
Df
n
/n (n = 3, 5, 7) returns to its preacid exposure value of about
28 Hz for all modes and the curves superimpose, correspond-
ingly, DD
n
also returns to its low preacid value, <1 10
6
. Thus,
after one acid challenge, the values of Df
n
/n and DD
n
, remain con-
stant. After 10 acid challenges, the values of Df
n
/n (n = 3, 5, 7) and
DD
n
return to their original values after each challenge. This indi-
cates that the mass of the cross-linked CH/AS surface as well as the
physical and chemical properties of the surface are the same post
acid treatment. In contrast, acid challenges of the physiosorbed
CH/AS surface resulted in a slow increase in Df
n
/n (n = 3, 5, 7)
and a decrease in DD
n
(Data not shown). Qualitatively, this result
indicates that during the acid challenges the mass of the physically
adsorbed CH/AS layer slowly decreases and the lm becomes more
elastic.
To calculate the hydrated thickness and shear modulus of each
successive layer, a Voigt based-viscoelastic model was used to t
the experimental data, Df
n
/n and DD
n
for n = 3, 5 and 7. Fig. 5a
shows that the viscoelastic model (dashed lines) is in excellent
agreement with the experimental data for n = 3, 5, and 7 (solid
lines). The thicknesses from the viscoelastic model are plotted ver-
sus time in Fig. 5b. The AS layer on HA has a thickness of 16 nm and
shear modulus of 9. 10
5
N/m
2
. After exposure to chitosan, the
mixed CH/AS layer has a thickness of 37 nm and the lowest shear
modulus, 2 10
5
N/m
2
. Upon rinsing the cross-linked chitosan
layer with DI water, the CH/AS layer thickness decreases further
to 10 nm and the shear modulus reaches its largest value,
2 10
6
N/m
2
. After 10 acid challenges the CH/AS layer thickness
remains 10 nm.
Fig. 6 compares the thickness of the cross-linked CH/AS and
physisorbed chitosan/AS layers after ve acid challenges. The data
in Fig. 6a reproduces the initial ve challenges in Fig. 5b. For the
physically adsorbed CH/AS surface, the acid exposure and DI water
rinse cycle is reduced to 4 min., about 2x less than in the cross-
linked CH/AS case. Fig. 6b shows that the physisorbed CH/AS layer
thickness decreases from 16 nm to 6 nm after the ve acid chal-
lenges. Thus, this layer appears less stable to each acid challenge,
decreasing by 2 nm per challenge, relative to the crosslinked CH/
AS. This difference indicates that the thiner, more elastic cross-
linked CH/AS layer resists desorption in acidic conditions relative
to the thicker, more viscous physisorbed CH/AS layer. Higher
charge density for the crosslinked case slows down permeability.
In summary, during the acid challenge, the cross-linked CH/AS
layer and corresponding HA layer coated on the QCM sensor show
no change in mass or thickness. However, the mass and thickness
of the physically adsorbed CH/AS layer and corresponding HA layer
coated on the QCM sensor decrease. AFM experiments were
performed to determine if the thinning of the CH/AS layer corre-
sponded to an erosion of the HA surfaces.
3.3. HA surface characterization and erosion using AFM
To evaluate the ability of both chitosan-modied AS layers to
prevent acid penetration and subsequent erosion of HA, the surface
morphology and roughness of each dry surface was characterized
using tapping mode AFM. Fig. 7 shows representative topography
images (1 1 lm
2
scan area) of the crystal used in the QCM-D
experiments that correspond to surfaces described in Fig. 3.
Fig. 7a is the image of the as-received nanocrystalline hydroxyap-
atite, HA, on the QCM sensor. The root mean square roughness
(R
rms
) of the HA surface is 2.2 0.2 nm. The polycrystalline HA do-
mains are circular and at with a diameter of 96 nm. The nano-
crystalline particle shapes are readily observed in the phase
image. This result agrees with the manufacturers characteristics,
which identify an RMS value of less than 3 nm. Fig. 7b shows the
surface topography of the physically adsorbed AS layer on HA sur-
face. The nanocrystalline domains of HA are still observed under
the AS layer. The QCM studies showed that the AS layer on the
HA surface is 16 nm and relatively stiff with a shear modulus of
9 10
5
N/m
2
. Thus, the AFM results are consistent with a thin,
stiff, conformal layer of AS over the nanocrystalline domains of
HA. Relative to the HA surface, the surface roughness,
2.4 0.2 nm, has not changed signicantly. Fig. 7c shows the
topography of the physically adsorbed chitosan on the AS surface.
Relative to the AS coated HA, the R
rms
decreases from 2.4 0.2 nm
to 0.4 0.1 nm consistent with a surface that is smooth and rather
featureless. QCM studies indicated that this layer is relatively thick,
37 nm, and exhibits the lowest shear modulus, 2 10
5
N/m
2
. The
nanocrystalline HA domains are no longer observed, suggesting
that the CH/HA layer lls in the valleys between the HA domains.
Upon exposure to an oxidizing agent that cross-links molecules,
the R
rms
increases to 1.5 0.3 nm (Fig. 7d). In addition to the
nanoscale features, larger features with a diameter of 200 nm
are observed. According to QCM-D studies, the cross-linked
Fig. 6. (a) Thickness of cross-linked chitosan on articial saliva after exposure to
citric acid for 5 min. followed by rinsing with water for 5 min. This sequence is
repeated 5x. (b) Thickness of physi-sorbed chitosan on articial saliva during
exposure to citric acid for 2 min. followed by rinsing with water for 2 min. Whereas
the physisorbed chitosan/AS mixture decreases from 16 nm to 6 nm, about 2 nm
per challenge, the cross-linked/AS mixture remains constant at 10 nm even though
the exposure times are about 2x that of the physisorbed system.
240 H.-S. Lee et al. / Journal of Colloid and Interface Science 385 (2012) 235243
CH/AS corresponds to a thin (10 nm), collapsed lm that is very
stiff, with the highest shear modulus, 2 10
6
N/m
2
. Thus the in-
crease in roughness observed in Fig. 7d is consistent with the prop-
erties of cross-linked CH/AS.
Fig. 7e shows the surface topography of cross-linked chitosan
after ten acid challenges. The R
rms
value (1.8 0.2 nm) did not dif-
fer signicantly from that of cross-linked CH/AS layer before the
acid challenges, which was consistent with QCM-D results. To
directly examine the HA surface, the cross-linked CH/AS layer
was removed by exposing the surface to an oxygen plasma. As
noted in the experimental section, this method was shown to re-
move the organic surface layer without damaging HA lm. Fig. 7f
shows the topography image after removing the cross-linked CH/
AS layer. Qualitatively, the surface features of the HA domains look
similar to those in Fig. 7a. Fig. 8 shows the R
rms
values for the plas-
ma treated physisorbed (left) and cross-linked (right) CH/AS layers
Fig. 7. AFM images of HA surface after various treatments. (a) HA surface alone. (b) AS layer on the HA surface. (c) Surface after exposure of AS-HA layer to chitosan. (d)
Surface after exposing chitosan to an oxidizing agent to cross-link the chitosan on the AS-HA. (e) Surface (d) after 10 min. exposure to citric acid, similar to the treatments in
the QCM-D experiments. (f) The HA surface (e) after removing the organic layer using an oxygen plasma. Note the similarity between the original HA surface and the
protected HA surface in Fig. 7a and f, respectively. The red lines represent line scans shown below each AFM image. RMS represents the root-mean-square roughness of the
corresponding surface in nm.
H.-S. Lee et al. / Journal of Colloid and Interface Science 385 (2012) 235243 241
on HA surface after acid challenges. For direct comparison, the
roughness of the bare HA before treatment is given. For the phys-
ically adsorbed CH/AS layer on HA surface after acid challenges, the
R
rms
value decreases from 2.2 0.2 nm to 1.6 0.1 nm, respec-
tively. This follows the QCM-D results, show a small decrease in
mass (about 2 nm per challenge), suggesting that acid can pene-
trate the physically adsorbed CH/AS layer and slightly erode the
HA. The bare HA surface and plasma treated cross-linked CH/AS
layer on HA surface after acid challenges, however, show similar
R
rms
values, 2.2 0.2 nm and 2.3 0.2 nm, respectively. Recall that
the QCM-D results show no mass change for the cross-linked CH/
AS layer on HA surface after acid challenges. Thus, the similar mor-
phologies of HA before and after indicate that the cross-linked CH/
AS layer protects the HA surface from acid.
4. Conclusions
We have found that both the cross-linked chitosan/articial sal-
iva surface and the physically adsorbed chitosan/articial saliva
surface served as a novel method to prevent acid erosion of the
model dental hydroxyapatite surface. Despite the ability of both
surfaces to decrease acid erosion however, the cross-linked chito-
san/articial saliva surface served as a slightly better surface for
the protection of HA. In this study, we have successfully developed
a novel protocol for the study of surfaces relevant to oral hard tis-
sue. We have also shown that QCMis a valid technique to study the
erosion of HA. It also would be benecial to examine the properties
of mixtures of chitosan and other known biocompatible polymers
for surface coatings. Our work, however, successfully lays the
groundwork for further exploration of HA erosion as monitored
by QCM.
Acknowledgments
This work received primary support from Colgate Palmolive
Company. HSL and RJC acknowledge support from the NIH Grant
RO1-HL60230. Partial support was also provided by the NSF/NSEC
(DMR08-32802) for facility use and NSF/Polymer Program
(DMR09-0793). We gratefully acknowledge use of QCM-D and ZP
equipment in the laboratory of Professor. D. Lee.
References
[1] C.M. Kreulen, A. Vant Spijker, J.M. Rodriguez, E.M. Bronkhorst, N.H.J. Creugers,
D.W. Bartlett, Caries Res. 44 (2010) 151.
[2] D.T. Zero, Eur. J. Oral Sci. 104 (1996) 162.
[3] H.M. Theuns, J.W.E. van Dijk, F.C.M. Driessens, A. Groeneveld, Arch. Oral Biol.
30 (1985) 37.
[4] A.K. Johansson, P. Lingstrom, T. Imfeld, D. Birkhed, Eur. J. Oral Sci. 112 (2004)
484.
[5] U. Lendenmann, J. Grogan, F.G. Oppenheim, Adv. Dent. Res. 14 (2000) 22.
[6] H.J. Guchelaar, A. Vermes, J.H. Meerwaldt, Support. Care Cancer 5 (1997) 281.
[7] M. Rillosi, G. Buckton, Pharm. Res. 12 (1995) 669.
[8] D. Patel, J.R. Smith, A.W. Smith, N. Grist, P. Barnett, J.D. Smart, Int. J. Pharm. 200
(2000) 271.
[9] P. Aksungur, A. Sungur, S. Unal, A.B. Iskit, C.A. Squier, S. Senel, J. Control.
Release 98 (2004) 269.
[10] D. Patel, A.W. Smith, N. Grist, P. Barnett, J.D. Smart, J. Control. Release 61
(1999) 175.
[11] O. Etienne, A. Schneider, C. Taddei, L. Richert, P. Schaaf, J.C. Voegel, C. Egles, C.
Picart, Biomacromolecules 6 (2005) 726.
[12] J.D.B. Featherstone, A. Lussi, Understanding the chemistry of dental erosion, in:
A. Lussi (Ed.), Dental Erosion-from Diagnosis to Therapy, Karger, Bern, 2006, p.
66.
[13] A.C. Magalhes, A. Wiegand, D. Rios, H.M. Honorio, M.A.R. Buzalaf, J. Appl. Oral
Sci. 17 (2009) 75.
[14] J. Vlacic, I.A. Meyers, L.J. Walsh, Aust. Dent. J. 52 (2007) 175.
[15] D. Rios, A.C. Magalhaes, M.A.D.M. Machado, S.M.B. da Silva, R.D.Z. Lizarelli, V.S.
Bagnato, A.R. Buzalaf, Photomed. Laser Surg. 27 (2009) 743.
[16] M. Beyer, J. Reichert, E. Heurich, K.D. Jandt, B.W. Sigusch, Dent. Mater. 26
(2010) 831.
[17] D. Churchley, G.D. Rees, E. Barbu, T.G. Nevell, J. Tsibouklis, Int. J. Pharm. 352
(2008) 44.
[18] L.H. Gracia, A. Brown, G.D. Rees, C.E. Fowler, J. Dent. 38 (2010) S4.
[19] R. Tarsi, R.A.A. Muzzarelli, C.A. Guzman, C. Pruzzo, J. Dent. Res. 76 (1997) 665.
[20] J.D. Bumgardner, R. Wiser, P.D. Gerard, P. Bergin, B. Chestnutt, M. Marini, V.
Ramsey, S.H. Elder, J.A. Gilbert, J. Biomater. Sci., Polym. Ed. 14 (2003) 423.
[21] S.Y. Chae, M.K. Jang, J.W. Nah, J. Control. Release 102 (2005) 383.
[22] I. Orienti, T. Cerchiara, B. Luppi, F. Bigucci, G. Zuccari, V. Zecchi, Int. J. Pharm.
238 (2002) 51.
[23] G.C. Ritthidej, T. Phaechamud, T. Koizumi, Int. J. Pharm. 232 (2002) 11.
[24] P. Tengamnuay, A. Sahamethapat, A. Sailasuta, A.K. Mitra, Int. J. Pharm. 197
(2000) 53.
[25] M.V. Shamov, S.Y. Bratskaya, V.A. Avramenko, J. Colloid Interface Sci. 249
(2002) 316.
[26] Q.S. Zhao, X.J. Cheng, Q.X. Ji, C.Z. Kang, X.G. Chen, J. Sol-Gel Sci. 50 (2009) 111.
[27] W. Senaratne, L. Andruzzi, C.K. Ober, Biomacromolecules 6 (2005) 2427.
[28] C. Whitehouse, J.Y. Fang, A. Aggeli, M. Bell, R. Brydson, C.W.G. Fishwick, J.R.
Henderson, C.M. Knobler, R.W. Owens, N.H. Thomson, D.A. Smith, N. Boden,
Angew. Chem. Int. Ed. 44 (2005) 1965.
[29] T.W. Kelley, P.A. Schorr, K.D. Johnson, M. Tirrell, C.D. Frisbie, Macromolecules
31 (1998) 4297.
[30] E.M. Blanco, M.A. Horton, P. Mesquida, Langmuir 24 (2008) 2284.
[31] N.A. Nikonenko, O.N. Tretinnikov, J. Appl. Spectrosc. 75 (2008) 878.
[32] J.R. Smith, M.T. Cicerone, C.W. Meuse, Langmuir 25 (2009) 4571.
[33] N. Sakamoto, T. Shioya, T. Serizawa, M. Akashi, Bioconjugate Chem. 10 (1999)
538.
[34] N.A. Nikonenko, I.A. Bushnak, J.L. Keddie, Appl. Spectrosc. 63 (2009) 889.
[35] T. Indest, J. Laine, V. Ribitsch, L.S. Johansson, K. Stana-Kleinschek, S. Strnad,
Biomacromolecules 9 (2008) 2207.
[36] T. Indest, J. Laine, L.S. Johansson, K. Stana-Kleinschek, S. Strnad, R. Dworczak, V.
Ribitsch, Biomacromolecules 10 (2009) 630.
[37] I.G. Sedeva, R. Fetzer, D. Fornasiero, J. Ralston, D.A. Beattie, J. Colloid Interface
Sci. 345 (2010) 417.
[38] G.V. Martins, E.G. Merino, J.F. Mano, N.M. Alves, Macromol. Biosci. 10 (2010)
1444.
[39] M. Rodahl, F. Hook, A. Krozer, P. Brzezinski, B. Kasemo, Rev. Sci. Instrum. 66
(1995) 3924.
[40] A. Naderi, J. Iruthayaraj, T. Pettersson, R. Makuska, P.M. Claessont, Langmuir 24
(2008) 6676.
[41] G.V. Martins, J.F. Mano, N.M. Alves, Carbohydr. Polym. 80 (2010) 570.
[42] P. Myllytie, J. Salmi, J. Laine, Bioresources 4 (2009) 1647.
[43] G. Sauerbrey, Z. Phys. 155 (1959) 206.
[44] M.D. Ward, D.A. Buttry, Science 249 (1990) 1000.
Fig. 8. A. RMS roughness of (I) initial HA surface and (II) the nal HA surface after removing either physisorbed CH/AS layer (green) or cross-linked CH/AS layer (purple) using
an oxygen-plasma. For (II) samples were challenged with 10 sequences of acid exposure/rinse. Whereas the roughness of the HA coated with the cross-linked CH/AS layer
remains 2.3 nm, the roughness of the HA surface treated with the physisorbed CH/AS layer decreases from 2.2 to 1.6 nm, suggesting that some erosion of the HA has occurred.
B. Cartoon illustrating how acid exposure to the HA surface can lead to a decrease in surface roughness. (For interpretation of the references to colour in this gure legend, the
reader is referred to the web version of this article.)
242 H.-S. Lee et al. / Journal of Colloid and Interface Science 385 (2012) 235243
[45] T.W. Schneider, D.A. Buttry, J. Am. Chem. Soc. 115 (1993) 12391.
[46] B.D. Vogt, E.K. Lin, W.L. Wu, C.C. White, J. Phys. Chem. B 108 (2004) 12685.
[47] B.D. Vogt, C.L. Soles, H.J. Lee, E.K. Lin, W.L. Wu, Langmuir 20 (2004) 1453.
[48] H. Lee, L.S. Penn, Macromolecules 41 (2008) 8124.
[49] H. Lee, L.S. Penn, Langmuir 25 (2009) 7983.
[50] H. Lee, D.M. Eckmann, D. Lee, N.J. Hickok, R.J. Composto, Langmuir 27 (2011)
12458.
[51] F. Hook, B. Kasemo, T. Nylander, C. Fant, K. Sott, H. Elwing, Anal. Chem. 73
(2001) 5796.
[52] F. Hook, M. Rodahl, P. Brzezinski, B. Kasemo, Langmuir 14 (1998) 729.
[53] T.J. Halthur, P.M. Claesson, U.M. Elofsson, Langmuir 22 (2006) 11065.
[54] S.J. Ban, Dent. Mater. J. 22 (2003) 467.
[55] H. Agheli, J. Malmstrm, P. Hanarp, D.S. Sutherland, Mater. Sci. Eng., C 26
(2006) 911.
[56] T. Sandberg, H. Blom, K.D. Caldwell, J. Biomed. Mater. Res. Part A 91A (2009)
773.
[57] T. Sandberg, M.K. Ott, J. Carlsson, A. Feiler, K.D. Caldwell, J. Biomed. Mater. Res.
Part A 91A (2009) 762.
[58] M. Cardenas, T. Arnebrant, R.K. Thomas, G. Fragneto, A. Rennie, L. Lindh,
Biomacromolecules 8 (2007) 65.
[59] C. Christersson, P. Glantz, The role of salivary lm formation in oral, in: H.J.
Busscher, L.V. Evans (Eds.), Health: Oral Biolms and Plaque Control, Harwood
Academic Publishers, Australia, 1998, p. 83.
[60] M. Hannig, A. Joiner, The structure, function and properties of the acquired
pellicle, in: R.M. Duckworth (Ed.), The Teeth and Their Environment, Karger,
Bern, 2006, p. 29.
[61] A.S. Perlin, Glycol-cleavage oxidation, in: W. Pigman, D. Horton (Eds.), The
Carbohydrates, Chemistry and Biochemistry, Academic, New York, 1980, p.
1167.
[62] B.H. Nicolet, L.A. Shinn, J. Am. Chem. Soc. 61 (1939) 1615.
[63] I.M.N. Vold, B.E. Christensen, Carbohydr. Res. 340 (2005) 679.
[64] A. Dedinaite, M. Lundin, L. Macakova, T. Auletta, Langmuir 21 (2005) 9502.
[65] J.D. Rudney, Adv. Dent. Res. 14 (2000) 29.
[66] A. Young, G. Smistad, J. Karlsen, G. Rlla, M. Rykke, Adv. Dent. Res. 11 (1997)
560.
[67] T. Pettersson, A. Dedinaite, J. Colloid Interface Sci. 324 (2008) 246.
H.-S. Lee et al. / Journal of Colloid and Interface Science 385 (2012) 235243 243