iir J Oral Sci 1995: 103: 322-326 Copyright Munksgaard 1995

Printed in Denmark , All rights reserved

EUROPE-APJ JOURNAL OF
ORAL SCIENCES
fSSN 0909-883

Influence of light intensity on A.J.Feilzer, L. H.Dooren, A. J.de Gee

and C. L. Davidson
Department of Dental Materials Science,

polymerization shrinkage and Academic Center for Dentistry Amsterdam

(ACTA), Amsterdam, the Netherlands

integrity of restoration-cavity
interface
Feibei- AJ, Dooren LH, de Gee AJ, Davidson CL: Influence of light intensity on
polymerization shrinkage and integrity of restoration-cavity interface. Eur J Oral
Sci 1995; 103: 322-326. Munksgaard, 1995.

The restJts of this study showed that the use of high intensity curing light units
negatively affected the integrity of the restoration-cavtty interface in class V re-
storations. This is explained by the high reaction rates of light curing resin
composites. The interfacial integrity was better preserved with low light intensity Key words: Adaptation; Setting stress: Flow;
as it extends the visco-elastic stage of the setting materials, thereby moderating Light intensfty; Conversion rate
the setting stress development. The ultimate polymerization shrinkages for both
A, J, Feilzer, ACTA, Department of Dental Mater-
conditions were equal, which suggested equal degrees of conversion and thus ials Science, Louwesweg 1, 1066 EA
equal material properties. The results may alleviate the trend in using higher Amsterdam, the Netherlands, Telefax: +31-20
intensity light curing units and in particular the development of units with laser 6692726, E-mail: a,feilzer(a;acla.ni
beams in an attempt to further increase conversion rates. Accepted for pubiicafion March 1995

A high degree of conversion, which is primarily
related to curing light intensity and exposure time,
is an important factor for the longevity of a resto-
ration. However, the higher the conversion in resin
composites, the higher the shrinkage (1). In adhe-
sive resin composite restorations, shrinltage brings
about stress, which frequently leads to marginal
leakage. Hence the objective of obtaining optimal
material properties of composite restorations con-
flicts Vfcfith the objective of achieving optimal conti-
nuity of integrity at the composite-cavity interface.
A well-accepted approach to reduce this stress
in direct resin composite restorations is the use of
elastic "stress absorbing" lining materials which
may render sufficient strain to replace to the stress
(2, 3). For heavily loaded restorations however, this
method may negatively affect the stability and lon-
gevity, as the restoration may flex in the low mod-
ule support. Another approach for polymerization
stress reduction which does not affect the final me-
chanical stability of the restoration can be found
in controlling the plasticity (flow capacity) of the
restoration during curing (4). Based on studies of
chemically initiated composites, it has been sug-
gested that a positive relation exists between set-
ting time and flow characteristics of resin compos-
ites (5). An increase in the setting time of light cur-
ing composites can be achieved by decreasing the
curing light intensity The question is to what ex-
tent the curing light intensity can be reduced with-
out sacriflcing the ultimate material properties.
The light intensity of commercially available light
units varies from approximately 50 to 900 mW/cm^
(6, 7). RuEGGEBERG et al (8) have shown that the
light intensities should be above 233 mW/cm' to
provide sufficient energy for an adequate cure of
composite layers of I mm thickness. So, at the
lowest acceptable value of 233 mW/cm^, and ap-
plying the composite in separately cured layers
with a maximum thickness of 1.0 mm, the ulti-
mate cure and thus the stiffness of the restoration
may not be influenced. Only with very low inten-
sity curing light units, even with prolonged light
exposure times, are problems with curing to be
expected.
The aim of this study was to determine the in-
fluence of the intensity of the curing light on the
integrity of the restoration-cavity interface of in
vitro Class V composite restorations and on the
polymerization shrinkage as a measure for the
 Light Intensity versus Integrity of Adhesion 323
Table 1
Materials used in this study
Brand Manufacturer Batch
Clearfil Liner Bond System Cavex Holland BV, Haarlem, N L 41119
- CA Conditioner
- SA Primer II
- Photo Bond III
- Protect Liner
Composite
- Clearfil Lustre (US) Cavex Holland BV 0012A

degree of conversion. In addition, the effect of a facturer's instructions: CA agent / SA primer /

stress absorbing elastic intermediate was investi-
gated.
Materials and methods
Determination of the integrity of the restoration-
cavity interface - The adhesive Clearfil Liner Bond
system in combination with the composite restor-
ative material Clearfil Lustre was used in this study
(Table 1). The Clearfil Liner Bond system consists
of a surface conditioner (CA agent); a primer (SA
primer); a bonding agent (Clearfil Photo Bond);
and an elastic liner (Protect Liner). The use of Pro-
tect Liner is advised by the manufacturer for all
clinical cases with a vital pulp. The CA agent is a
citric acid - calcium chloride based tooth surface
conditioner to be used for both dentin and enamel.
In 24 extracted human teeth (premolars and ca-
nines), V-shaped Class V cavities were prepared
(Fig. 1) in the labial and lingual surfaces with the
cervical margin approximately 1 mm below the
enamel-cement junction. Cavities were made with
a 109/010 diamond bur at high speed with water
coolant as uniform as possible with a depth of
1.50.1 mm in the dentin. The teeth were random-
ly divided into three groups. The groups differed
from each other with respect to the elasticity of
the materials used to fill the cavity. In group I the
restorative procedure was according to the manu-
Photo Bond / Protect Liner / Clearfil Lustre. In
group II, the teeth were restored as group I,, but
without the use of Protect Liner. In group III, the
cavities were filled as group I, but the composite
Clearfil Lustre was completely replaced by Protect
Liner. For each group, half of the samples were
cured with a Translux CL curing unit (Kulzer,
Wehrheim, Germany) at a light intensity of 250
mW/cm^ (measured with the Curing Radiometer,
Demetron Research, Danbury, CT, USA) and the
other half was cured with the Elipar II light curing
unit (ESPE, Seefeld, Germany) with an intensity
of 650 mW/cm^. The restorations were finished
using a finishing diamond bur and Sofex discs
(3M, St Paul, MN, USA). After thermocycling
(400 X, 15-65C) the specimens were sectioned
bucco-lingually through the restoration with a dia-
mond saw (Isomet, Buehler, Evanston, IL, USA)
at slow speed. Impressions were made with a poly-
vinylsiloxane impression material (Type I wash,,
KERR (Europe), Basel, Switzerland) after drying
the sectioned surface by way of wiping off the sur-
face with a tissue. This superficial drying was done
to avoid the risk of creating artifacts by dehydra-
tion. Araldite (Ciba-Geigy NV, Arnhem, the Neth-
erlands) casts were made which, after gold sputter-
ing (Sputter coater S150B; Crawley, Edwards, West
Sussex, UK) were examined in the Scanning
Electron Microscope (SEM) (Philips XL20, Eind-

Table 2
Integrity of the restoration-cavity interface (mean scores as percentage of perfect integrity)
Zone
1 9 3 4 5
Group I 250 mW/cm^ 100% 100% 100% 100% 100%
650 mW/cm^ 88% iOO% 69% 94% 100%
Group II 250 mW/cm^ 100% 100% 88% 100% 100%
650 mW/cm^ 100% 100% 69% 100% 94%
G r o u p III 250 mW/cm= 100% 100% 100% 100% 100%
650 mW/cm- 88% 100% 94% 100% 100%
Group I restorative procedure according to manufacturer: CA agent/SA primer/P'hoto Bond/Protect Liner/Clearfil Lustre restor-
ative.
Group II as groiip I, but without Protect Liner,
Group III as group 1, but instead of OeaTfll Lustre, Protect Liner was used as the restorative material.
 324 Eeilzer et al.
Eig. 1. Schematical representation of the experimental Class V
restoration and the division of the restoration-cavity interface
into the five zones in which the integrity was scored.
hoven, the Netherlands) at magnifications to
1000 X. To judge the extension of the integrity of
the restoration-cavity interface, the interface was
divided into five zones of equal length (Fig.l). The
interfacial condition in each zone was scored on a
two-way scale: "open" or "closed". For example, if
interfacial separation was observed in zone 1 and
in zones ?>-A, then zones 1, 3 and 4 scored "open" polymerization shrinkage determinations grad-
while zones 2 and 5 scored "closed". For each
group, mean scores per zone were calculated as
percentages of perfect integrity. Statistical compar-
ison of the results was done using a loglinear
model analysis with SPSS-pc package (9).
Determination of polymerization shrinkage - The
influence of light intensity on the polymerization
shrinkage of Clearfil Lustre was determined with a
linometer (10) on 1.50-mm-thick samples for light
intensities of 250 mW/cm- and 650 mW/cm^. The
linear polymerization contraction (lin%) was regis-
tered continuously with a computer for a period
of 60 min at 23 1C. The data were converted into
volumetric values (vol%=31in% - 0.03(hn%)2) (10).
For each light intensity, five measurements were
carried out and averaged, to obtain one mean
Table 3
Results of statistical analysis
l=zone (1 to 5)
2=group (1 to 3)
3=m'ft'/cm- ('250 mW/cm^' or '650 mW/cm^'
4=score ('open' or 'closed')
Model G^ df
#1 [123j[4] 52.92 23
#2 [123j[141 33.27 20
#3 [I23][24] 49.79 21
#4 [123][34] 38.75 22
Difference model #i and #2 G^= 19.65
Difference model #1 and #3 G^=3,13
Difference model #1 and #4 G^=14,17
shrinkage curve per intensity Statistical differences
between both conditions were revealed by way of a
r-test.
Results
Table 2 shows the mean scores as percentages of
perfect integrity of the restoration-cavity interface
from the SEM observations. Table 3 gives the re-
sults of the statistical analysis. On a P=0.05 level,
no significant effect of the restorative method
(Group I vs II vs III) was found. Inspection of the
parameter estimates showed that the restorations
cured with the low intensity light curing unit (250
mW/cm-), had a significantly lower amount of de-
fects at the interface than with the high intensity
unit (650 mW/cm-). No significant difference per
zone could be found. Closed zones (perfect integri-
ty) were significantly scored more than open zones
(interfacial separation).
Figs. 2 and 3 show the mean polymerization
shrinkage curves of Clearfil Lustre cured with 250
mW/cm' and 650 mW/cm- for the first 2 min and
for 60 min, respectively. The standard error of the
-6S0 mW/cm"
-250mW/cni=
time (mini
Fig, 2 Polymerization shrinkage (n = 5) in relation to curing
light intensity of Clearfil Lustre (1,5 mtn thickness) for the first
2 min of cure. The curves differ significantly frotn 0.05 min
onward (P<0,05).
- 6 5 0 mW/cm"
S - 2 5 0 mVltonf
30
lime (man)
Ei, 3. Polymerization shritikage (n=5) in relation to curing
Eig.
light intensity of Clearfil Lustre (1.5 mm thickness) for the first
hour of curing. The curves do not differ significantly frotn ap-
proximately 15 tninutes onvfard (P<0.05).

ually increased from 0.01 vol% for the early part
of the determination to 0.05 vol% for the later
part. The shrinkages of Clearfil Lustre resulting
from activation at 250 mW/cm^ and 650 mW/cm-
showed significant differences between 3 s and 15
min. Past 15 min, the shrinkage values became
equal.
Discussion
The bond strength of the latest generation dentin
bonding agents is often superior to the cohesive
strength of the substrate (11-13). However, mar-
ginal leakage studies with the same materials also
reported separation of some sort at the interface
(11, 13, 14). Apparently, adhesion of the restora-
tion to the entire cavity wall is not solved only by
strong bonding. Shrinkage stress and not inade-
quate bond strength is to blame for adhesive fail-
ures in composite resin restorations. The results of
the SEM observations (Table 2) showed, in gen-
eral, acceptable continuity of integrity of the resto-
ration-cavity interface for both intensities. The
number of defects per zone was low. However, the
integrity of the restorationcavity interface of the
restorations cured at low intensity was significantly
better than the integrity after curing at high inten-
sity. This agrees with the findings of UNO & As-
MUSSEN (4), who also found a significant improve-
ment of the marginal adaptation at reduced light
intensities. This can be explained by the difference
in polymerization rates resulting from the activa-
tion with different light intensities. During curing,
the visco-elastic behavior of the composite changes
from viscous to visco-elastic to elastic. This process
leads to stress development, as the restorative ma-
terial is adhesively bonded to the cavity walls. In
the viscous stage, no stress development will occur,
whereas stresses can partly be relieved by flow and
elastic strain in the visco-elastic stage and, finally,
by compliance of the surrounding structures. Too
high stress levels will eventually lead to interfacial
separation. As the rate of conversion determines
the rate of shrinkage stress development and the
ultimate stress level, any retardation of the reaction
will contribute to slowing down the shrinkage
stress development and its ultimate value. In addi-
tion, the time period for viscous flow will be ex-
tended. The polymerization rate of Clearfil Lustre,
as reflected by the shrinkage rate (Fig. 2), was
slowed down significantly when the curing light in-
tensity was reduced from 650 to 250 mW/cm^. The
main portion of stress reduction, favoring the in-
terfacial integrity, is expected to occur during the
first 10 s of light activation. During this period the
conversion is less than 50% of that obtained with
the high intensity unit (Fig. 2). The moderation of
Light Intensity versus Integrity of Adhesion 325
the reaction may also contribute to the overall
stress reduction by allowing more yielding of the
free surface of the restoration to the underlying
contracting bulk, as a result of a slower stiffness
development.
From Fig. 3 it can be seen that the shrinkage
curves of Clearfil Lustre resulting from activation
at 250 and 650 mW/cm^ reached the same level
within 1 h. This may illustrate (1) that in the 1.5
mm thick samples the same degree of conversion
could be obtained with both light intensities.
Therefore, it is assumed that the final mechanical
properties were not influenced in the particular res-
toration geometry. The reduced light intensity as
used by UNO & ASMUSSEN (4) did influence the me-
chanical properties, but this may be due to a light
intensity which is considerably lower than the 250
mW/cm^ used in this study.
The influence on interfacial integrity from the
elastic behavior of the various components of the
Liner Bond system follows from a comparison of
scores between groups I, II and III. Whereas in-
terfacial integrity of the restorations in groups I,
II and III was significantly influenced by influen-
cing the visco-elasticity during setting, no statisti-
cal difference was found by changing the elasticity
from a material choice. Group I comprised re-
storations with Protect Bond applied as an elastic
stress absorbing layer between cavity wall and
Clearfil Lustre. In group II, Clearfil Lustre was
used without Protect Bond, while in group III,
Protect Bond was used as the restorative instead of
Clearfil Lustre. An increase of elastic compliance
to setting stresses was expected in the order: group
I K group I< group III, but no influence on integ-
rity could be detected after statistical comparison
of the groups. Two main reasons may account for
the discrepancy with earlier findings that stress ab-
sorbers favor the cavity wall-restoration integrity
(2, 3). In the previous studies (2, 3) one of the early
bonding systems was used. The bond strength of
the present generation, which includes the Clearfll
Liner Bond system, has improved considerably. In
addition, the choice of Clearfil Lustre as the filling
material was somewhat unfortunate, as it subse-
quently appeared that its Young's module was of
the same order as that of Protect Bond (the 1-h
value of Protect Bond was 4.9 GPa and of Clearfil
Lustre 6,1 GPa,, measured in a 3-point bending set-
up). The use of a stiffer composite might have re-
vealed the beneficial effect of building in more elas-
ticity. To statistically express the small differences
from the present choice of materials, larger sample
sizes would have been required. However, it is
questionable whether these small differences will
be of any clinical relevance. A conclusion which
can be drawn for the particular cavity preparation.
326 Feilzer et al.
is that bulk filling with Clearfil Lustre without ap-
plying Protect Liner is as acceptable as the proce-
dure which includes the Protect Liner.
The results of this study show that the use of
light curing units with lower light intensities con-
tributed to a better integrity of the restoration-
cavity interface. As the degree of conversion is de-
pendent not only on the light intensity but also on
exposure time, type of resin, shade of composite,
depth of the restoration etc., the results of this
study should not be generalized and for the mo-
ment should only be applied to the Clearfil Lustre
composite together with the Clearfil Liner Bond
system in the given shallow Class V cavity. A study
of other types of composites is in progress. The
intention of this study was to draw attention to
the fact that manipulation of the setting time to
influence flow characteristics may be a useful tool
to moderate the development of shrinkage stresses
for better restoration-cavity wall interface. This
should moderate the trend towards using higher
intensity curing units and also the development of
curing units with laser beams. Aiming for restora-
tions with the highest degree of conversion and
maximal mechanical properties interferes with ob-
taining optimal integrity of the restoration-cavity
interface.
It can also be seen from the present study that
it should be compulsory to report the curing light
intensity used in reports on the quality of compo-
site restorations.
Acknowledgment - The authors wish to acknowledge Dr. J. G.
C. VERHEY for the performance of the statistical analyses.
References
1. VENHO\'EN B A M , D E GEE AJ. DAVIDSON CL. Polymeriza-
tion contraction and conversion of light-curing BisGMA-
based methacrylate Resins. Biomaterials 1993; 14: 871-875.
2. KEMP-SCHOLTE C M , DAVIDSON CL, Marginal integrity re-
lated to bond strength and strain capacity of composite
resin restorative systems. J Prosthet Dent 1990; 64- 658-
664.
3. KEMP-SCHOLTE C M , DAVIDSON CL. Complete marginal
seal of Class V resin composite restorations effected by
increased flexibility. J Dent Res 1990; 69': 1240-1243.
4. UNO S. ASMUSSEN A. Marginal adaptation of a restorative
resin polymerized at reduced rate. Scand J Dent Res 1991;
99: 440^44.
5. FEILZER AJ, DE GEE AJ, DAVIDSON CL. Quantitative deter-
mination of stress reduction by flow in composite restora-
tions. Dent Mater 1990; 6: 167-171.
6. ELLINGSON O L , LANDRY RJ, BOSTRO.-H R G , .\n evaluation
of optical radiation emissions from dental visible photo-
polymerization devices. JADA 1986; 112: 67-70.
7. FAN PL, WOZNIAK W T , REYES W D . STANFORD JW Irradi-
atice of visible light-curing units and voltage variation ef-
fects. JADA 1987: 115: 442^(45.
8. RtjEGGEBERG FA, CAUGHMAN W F , CURTIS JW. DAVIS H C .
Factors affecting cure and depths within iight- activated
resin composites. Am J Dent 1993: 6: 91-9.
9. FlENBERG SE. The Analysis of cross-classified categorial
data. 2nd edn. Cambridge, MA: MIT Press, 1980.
10. D E GEE AJ, FEILZER AJ, DAVIDSON CL. Ti-ue linear poly-
merization shrinkage of unfilled resins and composites de-
termined with a linometer. Dent Mater 1993: 9: 11-14.
11. PKMI C, NUCCI C , DAVIDSON CL, MONTANARI G . Early
marginal leakage and shear bond strength of adhesive re-
storative systems. Dent Mater 1990; 6: 195-200.
13. 0ILO G, AuSTRHEiM EK. In vitro quality testing of dentin
adhesives. Acta Odontol Seand 1993; 51: 263-26"^9.
13. RETIEE H , MANDRAS RS, RUSSELL CM. Shear bond
strength required to prevent microleakage at the dentin/
restoration interface. Am J Dent 1994; 7: 4 3 ^ 6 .
14. .ABDALLAH A L DAVIDSON CL. Shear bond strength and
microleakage of new dentin bonding systems. Am J Dent
1993; 6: 295-298,