J Mater Sci: Mater Med (2009) 20:10011007

DOI 10.1007/s10856-008-3657-1

Effects of different preparation procedures during tooth

whitening on enamel bonding
Dustin Wilson Changqi Xu Liang Hong
Yong Wang

Received: 11 April 2008 / Accepted: 1 December 2008 / Published online: 13 December 2008
Springer Science+Business Media, LLC 2008

Abstract The objective of this study was to assess effects

of some clinically related preparation procedures during
tooth whitening on enamel bonding properties. Sixty-two
extracted human teeth were cleaned and divided into four
groups. Forty-two of the teeth were left with their natural
surface intact while 20 teeth were polished to form a flat
surface. Half of the tooth served as the experimental side
and received one of the two whitening products: Opalescence (10% carbamide peroxide) and Crest Whitestrips
(6.5% hydrogen peroxide), for 2 weeks. Post-bleaching
intervals included: 1 day, 1 week, and 2 weeks. On these
days, tooth (10 mm 9 1.5 mm 9 1.5 mm) sections were
evaluated using Raman spectroscopy, scanning electron
microscopy and tensile bond strength tests. T-test, ANOVA
test, and mixed model regression analysis were used to
assess the differences. No significant difference existed
between natural surface and polished surface teeth for all
groups at both Day One and Week Two (P [ 0.05). On
Day One, both treated groups had significant lower bond
strength than the control group (P = 0.002). After
2 weeks, no significant difference existed between any
group (P = 0.381). SEM indicated that resinenamel
interfaces in bleached enamel exhibited more defects in
granular formations when compared to the control. Raman
results indicated a lower degree of polymerization (DP) of
adhesive at the interface for treated teeth surfaces. In
summary, pre-bleaching surface treatments such as polish
or non-polish, had no effect on bond strength. Bleaching
significantly decreased bond strength initially, but after

D. Wilson
C. Xu
L. Hong
Y. Wang (&)

Department of Oral Biology, University of Missouri-Kansas City
School of Dentistry, 650 E 25th St, Kansas City,
MO 64108, USA
e-mail: [email protected]

2 weeks, bleaching had no significant effect on bond

strength. Storage time had significant effect on Opalescence treated enamel, but not on control and Whitestrip
treated enamel. The decrease of bond strength may be
related to interfacial defects and low DP due to oxygen
release after bleaching.

1 Introduction
At-home teeth whitening systems have become popular
and effective way to remove both intrinsic and extrinsic
stains from teeth. Many whitening products are available
on the market or through a dental professional for at-home
use. There are two forms of whitening products available to
the public, at-home bleaching products and in-office
bleaching products. To enhance the effects of the bleaching
process carbopol is added to most in office bleaching
agents. Carbopol is an additive in carbamide peroxide that
enhances the materials adhesion to enamel and extends the
release of oxygen from peroxide [1].
The whitening agent in most whitening systems is
hydrogen peroxide or carbamide peroxide which eventually converts to hydrogen peroxide and urea. Tooth
discoloration is thought to be removed by hydrogen peroxide by means of oxidation [2]. Whitening materials cause
changes in the morphology of enamel similar to etching
that include a loss of prismatic form [3]. These changes to
the surface of enamel are important to understanding what
occurs during the bonding of composite to enamel. Studies
have shown that for a period of up to 2 weeks after
bleaching has been completed, the bond strength of the
enamel surface to composite resin has a slight decrease
[46]. However, it is argued that all the teeth were polished

123

1002

to obtained flat enamel surface for bond strength studies,

which might not simulate the clinical situation where natural surfaced teeth are used. In addition, the reason for
reduction in bond strength of the composite resin for
2 weeks after whitening is not known yet. During the
bonding process of the adhesive and composite resin to the
enamel the residue from the whitening agent or the change
in surface morphology could be the cause of this bond
strength reduction. Studies have shown that if the surface
layer of enamel is removed prior to bonding then there is
no reduction in bond strength [5]. However dentists try to
conserve as much healthy tooth structure as possible when
performing restorations, so this is not a practical procedure
for clinicians to perform during restorations. It is important
to understand the change that occurs to enamel chemically
and/or morphologically to cause this reduction so that a
procedure can be developed for clinicians to eliminate or
shorten the period of waiting 2 weeks to perform enamel
surface bonding after teeth whitening. The goal of this
study was to understand why there is a reduction in bond
strength for 2 weeks after whitening has been completed
by analyzing the enamel surface chemically and morphologically and to assess if there is a difference between
polished surfaced and natural surfaced teeth for tensile
bond strength tests.

2 Materials and methods

2.1 Specimen preparation
Sixty-two extracted human teeth with no surface caries or
enamel defects were stored at 4C in sterile Delbeccos
phosphate saline (cellgro DPBS, Mediatech, Inc., Herndon,
VA,USA) with 0.002% sodium azide were used for this
study. The teeth were collected after informed consent was
obtained under a protocol approved by the UMKC adult
health sciences institutional review board. At the start of the
study the teeth were cleaned with pumice, water, and a soft
rubber prophy cup on a slow speed hand-piece. The teeth
were then stored in an artificial saliva solution for the
remainder of the study at 37C. To prepare 1 L of artificial
saliva 0.22 g CaCl2 2H2O, 0.12 g KH2PO4, 11.16 g KCL
and distilled water were used. The pH value of the solution
ranged from 7.04 to 7.07 and the solution was changed every
3 days. The pH of the solution was measured using the
Accumet pH meter (Fisher Scientific, Pittsburgh, PA, USA).
2.2 Specimen selection and time intervals
Forty-two of the teeth with the natural surface left intact
were randomly divided into two groups. The remaining 20
teeth were polished to create a flat enamel surface using

123

J Mater Sci: Mater Med (2009) 20:10011007

320, 600, and 1200 wet grit CarbiMet Discs sandpaper

(Buehler Ltd, Lake Bluff, IL, USA) and were randomly
divided into two groups. All the teeth were sectioned in half
longitudinally. Half of the tooth served as the experimental
side and received one of the two whitening products
according to manufacturers directions for 2 weeks.
Twenty-one of the 42 teeth were randomly selected to
receive Crest Whitestrips (6.5% hydrogen peroxide)
(Procter& Gamble Company, Cincinnati, OH, USA) and
the other 21 teeth received Opalescence Gel (10% carbamide peroxide) (Ultradent Products, Inc., South Jordan,
Utah, USA). Ten of the 20 polished surfaced teeth were
randomly selected and received Crest Whitestrips (6.5%
hydrogen peroxide) and the other 10 polished teeth received
Opalescence Gel (10% carbamide peroxide). The Opalescence Gel was applied for 6 h per day and the Crest
Whitestrips were applied for 30 min twice a day for
2 weeks to the experimental side of the tooth. The other side
of the tooth did not receive any whitening product and
served as the control side. During the whitening process, the
teeth were stored in an environmentally controlled chamber
simulating oral cavity conditions (PLASLABS, Lansing,
Mich, USA). After whitening each day the teeth were
cleaned with distilled water and a tooth brush. The teeth
were then placed into artificial saliva solution at 37C.
Post-bleaching time intervals included: 1 day, 1 week,
and 2 weeks for the natural surface teeth and for the polished teeth the time intervals included: 1 day and 2 weeks.
For the teeth with their natural surface left intact, seven of
the teeth whitened with Crest Whitestrips and seven of the
teeth whitened with Opalescence where randomly selected
on 1 day, 1 week, and 2 weeks. For the polished surface
teeth, five teeth whitened with Crest Whitestrips and five of
the teeth whitened with Opalescence where randomly
selected on 1 day and 2 weeks. On these days, the teeth
were sectioned in half and the roots were removed with a
water-cooled diamond saw (ISOMET 1000, Buehler Ltd,
Lake Bluff, IL, USA).
2.3 Application of the bonding agent and composite
The bonding was completed using the adhesive, Single
Bond Plus (3 M ESPE, St. Paul, MN, USA), and composite
resin, Filtek Z250 (3 M ESPE, St. Paul, MN, USA). The
application of the adhesive started with etching the tooth
for 15 s then rinsing the tooth with distilled water for 10 s.
Next the specimen was dried with cotton and air dried for
10 s. The adhesive was then applied in two coats, the first
coat was air dried for 5 s and the second coat was air dried
for 10 s. Next the adhesive was cured for 10 s with visiblelight source (Dentsply Spectrum 800, Milford, DE, USA)
at 800 mW/cm2. The composite resin was then applied in
layers of no more than 2 mm height to a height of 5 mm

J Mater Sci: Mater Med (2009) 20:10011007

and then cured for 20 s. The specimens were then stored in

distilled water at room temperature for 24 h.
2.4 Mechanical testing regimen
Rectangle bar specimens (10 mm 9 1.5 mm 9 1.5 mm)
sectioned using a water-cooled diamond saw were used to
test mechanical properties. 5 mm of the length of each
specimen consisted of composite and the other 5 mm
consisted of tooth. Due to the difficulty of cutting beams
exactly 1.5 mm by 1.5 mm, the cross-sectional area of the
beam was calculated based on the exact dimensions measured by an electronic digital caliper (Marathon, Ontario,
Canada) before testing. The tensile properties were determined for all specimens from 24 to 48 h of bonding after
being stored in distilled water at room temperature. Specimens were tightly and fully attached to the upper and
lower grips using cyanoacrylate glue (Zapit, Dental Ventures of America, Corona, CA, USA) with the compositeenamel interface located in the middle of the 3.5 mm gap
between the grips and were load at a cross-head speed of
0.5 mm/min using SSTM-500 mechanical tester (United
Calibration Corporation, Huntington Beach, CA) with a
250 lb load cell. The ultimate tensile strength (UTS, MPa)
of each specimen was calculated as the maximum force at
the point of failure divided by the specimen cross-sectional
area. Prior to mechanical testing, the specimens were
carefully examined for defects. Specimens with defects
were not used. ANOVA, t-tests, and Duncan tests were
used to assess the differences between the natural surface
and polished surface groups, non-treated versus treated
groups, and time intervals.
2.5 Scanning electron microscopy (SEM)
The specimens were first viewed using SEM under their
natural surface without coating to preserve the interface for
micro-Raman spectroscopy. The dry specimens were cut
into 23 mm long bars with a diamond saw and mounted to
aluminum stubs using cyanoacrylate glue. The specimens
natural surface interface was analyzed with an XL30
ESEM-FEG SEM Microscope (FEI Company, Hillsboro,
OR, USA) at 500 V accelerating voltage to evaluate the
presence of the interface and resin tags. The specimens
were then examined using micro-Raman spectroscopy,
since this is a non-destructive technique, these specimens
were still available to be coated and analyzed using SEM.
Following micro-Raman analysis, the specimens were
coated with goldpalladium to prepare for SEM. Specimens were examined at a variety of magnifications with an
XL30 ESEM-FEG Microscope at 15 Kv accelerating
voltage. Analysis was performed on the adhesive interfaces
along with composite and enamel surfaces.

1003

2.6 Micro-Raman spectroscopy

The micro-Raman spectrometer (Lab RAMHR800, Horiba
Jobin Yvon, France) consisted of a laser beam (632 nm)
focused through both 509 and 1009 Olympus MPLAN
objectives to a 12 l beam diameter. Raman back-scattered light was collected through the objective and resolved
with a monochromator. The spectra were recorded. The slit
width of the spectrograph was set at 100 lm and the hole
diameter was set at 400 lm, providing spectral resolution
of 8 cm-1. Two scans of spectra (with a 60 s accumulation
time each) were obtained from each site. The laser power
was approximately 7 mW; no thermal damage of the
specimen was observed during measurement. An imaging
system and high-resolution monitor were used to allow for
visual identification of the position at which the Raman
spectrum was obtained. Each adhesive/enamel/composite
interface slab was mounted at the focus of the objective. To
investigate difference of the adhesive interface for specimens in all groups from Day One, Raman spectrum from
the adhesive the interface were collected The peak located
at 1,637 cm-1 indicates un-reacted C=C double bond of
the adhesive, and the peak in 1,608 cm-1 represents the
carbon to carbon bonds in aromatic rings in the Bis-GMA
molecules of the composite [7]. The ratios of the area at
peak 1,637 and 1,608 cm-1 for both the adhesive monomer
and polymer using the following equations: RPOLY-1
-1
in polymerized
MER = Area 1,637 cm /Area 1,608 cm
adhesive, and RMONOMER = Area 1,637 cm-1/Area
1,608 cm-1 in adhesive monomer before light curing.
The degree of polymerization, (PD), of the adhesive could
then be calculated using the following equation: PD =
1 - ((RPOLYMER)/(RMONOMER)).
The data were entered into and analyzed with SPSS 15.0
Statistical program (SPSS Inc., Chicago, IL 60606). All
individual pair-wise comparison was conducted either by ttest or One-Way ANOVA for match design. A mixed
model for multivariable regression analysis was developed
with time (1 day, 7 days, and 14 days) and treatment
(none, Opalescence, and Crest Whitestrips) as fixed effects
and tooth surface type (natural or polished) as a random
effect. The outcome variable was bond strength. Statistical
significance level was set at P = 0.05.

3 Results
3.1 Mechanical testing
The analysis of natural surface versus polished surface
tensile bond strengths for Day One and Week Two is
presented in Table 1. ANOVA test showed no significant
difference existed between natural surface and polished

123

1004

J Mater Sci: Mater Med (2009) 20:10011007

differences existed for the Control group and the Opalescence group. The significant difference for the control
group occurred between Day One compared to both Week
One and Week Two time intervals. The significant difference for the Opalescence group occurred between Week
One and Week Two time intervals.
Table 4 summarized locus of failure. Overall, composite
was the most common site for facture, followed by cohesive layer and interface. These three sites accounted for the
vast majority of fracture locus (81%).
Using a mixed model with time and treatment as fixed
effects and tooth surface type as a random effect, multivariable regression analysis showed that treatment
(P = 0.25) and time (P = 0.29) were not a significant
factor for bond strength after controlling for other factors
(type III test).

surface teeth for all groups at both Day One and Week Two
(P [ 0.05).
With no significant difference revealed, natural surface
groups and polished surface groups in the same time
interval and surface treatment were combined for analysis
of non-treated surfaces and treated surfaces. Also, natural
surface groups and polished surface groups within the same
surface treatment were combined for analysis of time
interval on tensile bond strength. Table 2 shows the results
for each time interval based comparing non-treated and
treated teeth surfaces. On Day One, ANOVA test revealed
a significant difference existed between the control group
and the treated surface groups for both Crest Whitestrip
and Opalescence (P = 0.002). For Week One time interval, a significant difference existed between the control
group and the Opalescence group but no significant difference existed between the control group and the Crest
Whitestrip group. For Week Two time interval, no significant difference existed between the control group and both
treated groups (P = 0.381).
Table 3 shows the results of each treatment group
compared to the three time intervals. No significant difference existed for the Crest Whitestrip group at any time
interval. According to Duncan Grouping, significant
Table 1 Day One and Week
Two results: natural surface
versus polished surface

3.2 SEM
All the representative SEM images shown are coated with
goldpalladium from Day One and the accelerating voltage
was 15 Kv. Figure 1a shows a representative SEM image
(magnification of 509) from the fractured surface of a
control group specimen. The fracture occurred mostly in

Group

1-Control

Day One
Week Two

2-Crest

Day One
Week Two

3-Opalescence

Day One
Week Two

Table 2 Results for all time

intervals based on surface
treatment

Bond strength (SD)

t-test

P value

0.855

0.401

1.362

0.180

0.825

0.419

0.372

0.714

0.288

0.777

2.070

0.053

11

Natural

28.4 (10.9)

Polish

25.2 (6.1)

14

Natural

18.6 (8.4)

10

Polish

21.8 (7.2)

Natural

21.7 (7.9)

Polish

19.3 (4.0)

Natural

20.5 (4.6)

Polish

21.2 (4.4)

Natural

20.0 (5.8)

Polish

19.3 (4.7)

Natural

25.2 (4.1)

Polish

20.0 (6.9)

Time

Group

Bond strength (SD)

ANOVA test

Duncan grouping

Day 1

1-Control

18

27.0 (9.1)

F = 6.897

2-Crest

10

20.7 (6.6)

P = 0.002

3-Opalescence

11

19.7 (5.3)

1-Control

14

22.6 (7.7)

F = 2.705

21.3 (7.5)

P = 0.077

AB

Week 1

2-Crest
3-Opalescence
Week 2

123

Surface

17.3 (4.7)

1-Control

24

22.9 (6.5)

F = 0.978

2-Crest

12

20.8 (4.4)

P = 0.381

3-Opalescence

12

22.6 (6.2)

J Mater Sci: Mater Med (2009) 20:10011007

Table 3 Results based on time
comparison within the same
surface treatment group

1005

Group

Time

Bond strength (SD)

ANOVA

Duncan grouping

1-Control

Day 1

27.0 (9.1)

F = 2.682

Week 1

22.6 (7.7)

P = 0.074

Week 2

22.9 (6.5)

Day 1

20.7 (6.6)

F = 0.044

Week 1

21.3 (7.5)

P = 0.957

Week 2

20.8 (4.4)

Day 1

19.7 (5.3)

F = 4.017

AB

Week 1

17.3 (4.7)

P = 0.024

Week 2

22.6 (6.2)

2-Crest

3-Opalescence

Table 4 Summary of locus of failure

Break point

1-Control
(%)

2-Crest
(%)

3-Opalescence
(%)

Total
(%)

Enamel

11

14

Interface

11

32

31

22

Cohesive

37

21

19

28

Composite

34

26

31

31

Damage in
removal

10

11

100

100

100

100

Total

10

the composite and only slightly through the adhesive

interface, revealing a strong adhesive layer. Figure 1b
shows a magnified view (magnification of 5009) of the
same control specimen as Fig. 1a at the adhesive interface
showing a high degree of consistency. Figure 2a, is an
image of a Crest Whitestrip specimen with magnification
of 509, a magnified view (magnification of 5009) of the
fractured surface of the same Crest Whitestrip specimen at
the adhesive interface is shown in Fig. 2b. The bubbling
effect of the adhesive layer in this image shows a lesser
regularity of the adhesive layer when compared to the

control specimen. Similarly, Fig. 3a shows a representative

image of fracture surface from the Opalescence group, a
magnified view of the same Opalescence specimen as
Fig. 3a at the adhesive interface is shown in Fig. 3b. As
with the Crest Whitestrip specimen, the Opalescence
specimen shows bubbling and a lesser degree of regularity
of the adhesive layer.
3.3 Micro-Raman spectroscopy
Representative Raman spectra obtained from bleached
groups and controls are shown in Fig. 4. The ratio
of the polymer, RPOLYMER = (Area 1,637 cm-1/Area
1,608 cm-1) for the control group equaled 0.205 and for
the treated surface group equaled 0.302. The higher the
ratio of the polymer is, the higher the amount of carbon
carbon double bonds that have not been polymerized
and therefore, a lower degree of polymerization of
the adhesive. The degree of polymerization, PD =
1 - ((RPOLYMER)/(RMONOMER)), calculated for the control
group equaled 0.947 and for the bleach-treated specimen
equaled 0.917. These results show a higher degree of
polymerization for the control group compared to the
treated surface group of the adhesive layer.

Fig. 1 Representative SEM

images of the fractured surface
from control group at Day One,
a magnification = 509; b
magnification = 5009

123

1006

J Mater Sci: Mater Med (2009) 20:10011007

Fig. 2 Representative SEM

images of the fractured surface
from Crest Whitestrip group at
Day One, a magnification =
509; b magnification = 5009

Fig. 3 Representative SEM

images of the fractured surface
from Opalescence group at Day
One, a magnification = 509;
b magnification = 5009

Fig. 4 Representative Raman spectra of adhesive resins in interfaces

between enamel and composite. a from control group; b from
Opalescence treated surface group

4 Discussion
No significant difference was found between natural surface teeth and polished surface teeth within the same group
and time interval. Previous studies have only used either
polished surface teeth or natural surface teeth and the
reliability of comparing these studies was unknown. These
results show that results from polished surface teeth studies
can be compared with studies with the natural surface left
intact for enamel bonding. This will be valuable to investigate studies using different concentrations of carbamide
peroxide and hydrogen peroxide along with different time
intervals of testing bond strength.

123

It was reported that bonding to enamel after application

of a whitening product had a lower bond strength than
bonding to a non-treated enamel surface, which is consistent with the data presented [4, 5]. Both of the treated
surface groups on Day One have significantly lower UTS
than the non-treated surface group. At the 2 week time
interval, there was no significant difference between the
control group and both the treated surface groups. It is
therefore important to understand the changes to the
enamel surface over this time period to understand this
reduction in bond strength immediately after whitening has
been completed.
The slight reduction in bond strength of the control
group from Day One compared to both Week One and
Week Two (Table 2) was noticed. Even though it is not
considered significant through the ANOVA test, it was
found to be relevant through the Duncan Grouping, and
should be noted. The slight difference between the two
time intervals shows that there is some factor acting on the
specimens during the storage intervals. Before this study,
data were lacking on the time effect on each individual
group. Possibly phosphate buffered saline (PBS) and/or the
artificial saliva have some effect on the enamel surface that
is exhibited over the storage interval. More research is
needed to understand what is causing this reduction in bond
strength for the control group.
The SEM analysis of the specimens from Day One
revealed that most of the fractures for the control group

J Mater Sci: Mater Med (2009) 20:10011007

occurred outside of the adhesive interface and only a small

portion of the fracture in the adhesive layer. These fractures
either occurred in the composite, enamel, or a combination
of the both. Fractures from the treated specimens showed a
greater proportion of the break occurring in the adhesive
layer rather through the composite or enamel layers. These
results reveal a stronger adhesive interface for the control
group when compared to treated specimens on Day One
which correlates well with the UTS data from Day One.
The SEM analysis of specimens in which the fracture
occurred through some part of the adhesive layer revealed
structural changes of the adhesive in treated surface specimens as compared to those of control surface specimens.
A granular or bubbling effect was seen in the adhesive
layer of the treated surface specimens. This bubbling did
not occur in the control specimens, which indicates a
change has occurred in the enamel surface during whitening for the treated surface specimens. Chemical analysis
through Raman spectroscopy showed a lower degree of
polymerization of the adhesive layer for the treated surface
specimens compared to the control specimens. As known,
the mechanical properties of polymer materials depend on
their degree of polymerization. The lower degree of polymerization contributes to the lower bond strength for the
treated specimens. To explain the lower degree of polymerization it is important to note that oxygen inhibits
polymerization of the adhesive. Oxygen is released from
carbamide peroxide and hydrogen peroxide during the
whitening process as follows [8]:
H2 NCONH2
H2 O2 ! H2 NCONH2 H2 O2
H2 O2 ! 2HO
HO H2 O2 ! H2 O HO2
HO2 ! H O
2
2 H2 O2 $ 2H2 O 2fOg $ 2H2 O O2

1007

The lower degree of polymerization can be contributed to

the oxygen release from the whitening agents.
The future goal of the study is to find an alternative to
delayed bonding, i.e., anti-oxidants, to reverse negative
effects of bleaching on bond strength, especially when
immediate bonding must be performed after bleaching.
This will save both the patient and dentist time and money.

5 Conclusions
Pre-bleaching surface treatments such as polish or nonpolish, had no effect on bond strength. Bleaching significantly decreased bond strength at day 1, but after 2 weeks,
bleaching had no significant effect on bond strength. SEM
results indicated that resinenamel interfaces in bleached
enamel exhibited more defects in the granular or bubblelike forms. Raman results indicated oxygen released from
bleach-treated enamel inhibited resin polymerization,
caused defects in interfaces and lowered bond strengths.

References
1. V.B. Haywood, Esthet Dent Update 2, 63 (1991)
2. R.E. Golstein, D.A. Garber, Complete Dental Bleaching (Quintessence Publishing, Chicago, 1995), pp. 165185
3. V.B. Haywood, H.O. Heymann, Quintessence Int 22, 515 (1991)
4. V. Cavalli, A.F. Reis, M. Giannini, G.M. Ambrosano, Oper Dent
26, 597 (2001)
5. E. Cvitko, G.E. Denehy, E.J. Swift Jr., J.A. Pires, J Esthet Dent 3,
100 (1991). doi:10.1111/j.1708-8240.1991.tb00976.x
6. M.V. Dishman, D.A. Covey, L.W. Baughan, Dent Mater 9, 33
(1994). doi:10.1016/0109-5641(94)90019-1
7. Y. Wang, P. Spencer, J Biomed Mater Res 59, 46 (2002). doi:
10.1002/jbm.1215
8. J.E. Dahl, U. Pallesen, Crit Rev Oral Biol Med 14, 292 (2003)

123