Light shielding effect of overlaying resin compositeon the photopolymerization cure kinetics of a resincomposite and a dentin adhesive

Antonio Apicellaa,*, Michele Simeoneb, Raffaella Aversaa,Alessandro Lanzac, Davide Apicellaa

aCRIB Centro di Ricerca Interdipartimentale sui Biomateriali, University of Naples, ‘Federico II’,Naples, ItalybDepartment of Operative Dentistry, School of Dentistry, University of Naples, ‘Federico II’, Naples, ItalycBiological Technologies Applied to Odontostomatologic Science School, Second University of Naples,‘SUN’, Naples, Italy

Received 28 October 2003; received in revised form 1 September 2004; accepted 29 September 2004

01do

EnDifa

KEYWORDSLight cure;Photo-polymerization;Shielding;DSC;Polymerization kinetic

09-5641/$ - see front matter Q 2005i:10.1016/j.dental.2004.09.012

* Corresponding author. Address: Degineering Production, University ofocleziano 328, 80122 Naples, Italyx: C39 81 7629103.E-mail address: apicella@unina.it (

Summary Objectives: The purpose of this study was to simultaneously determinethe impact of exposure times and incremental resin composite overlaying thicknesson the cure kinetics of a light activated composite and a dentin adhesive at selecteddepths of a simulated restoration.Methods: Levels and kinetics of polymerization of a light activated resin composite(Z250, 3M-ESPE) and dentin adhesive (Excite, Ivoclar) cured with a halogen light unit(Demetron, Kerr, USA) operating at low/medium intensity (500 mW/cm2) fordifferent exposure durations (20 and 60 s) were measured at selected depths (0.3,0.6 and 1 mm) using a modified calorimetric analyzer (DSC 25, METLLER-TOLEDO).Results: Final polymerization levels of materials up to 1 mm through the compositeare not significantly different while the light shielding effect of incremental resincomposite overlaying progressively reduces their reaction rates.Significance: Prolonged irradiation time at low/medium energies is effective forproper conversion of a resin composite and dentin adhesive; 60 s irradiation timeprovides the maximum obtainable conversion through the dental resin composite forthicknesses up to 1 mm.Q 2005 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

Academy of Dental Materials

partment of Materials andNapoli ‘Federico II’, via

. Tel.: C39 81 7629102;

A. Apicella).

Introduction

Replacement of mercury-containing dental fillingmaterials with resin-based composite ones isbecoming a reality in today’s dentistry, especially

Dental Materials (2005) 21, 954–961

www.intl.elsevierhealth.com/journals/dema

. Published by Elsevier Ltd. All rights reserved.

Light shielding effect of overlaying 955

in direct posterior medium size dental restorations[1]. Resin-based dental composites, formed ofmethacrylates containing glass fillers at a volumefraction of 0.5–0.6 [2] that, due to their by partialtransparency toward light, can be polymerized byadding light activated initiators to the system.Unfortunately, polymerization of resin-basedcomposites is associated with both shrinkage andstiffening. Under high intensity light-cure con-ditions, almost instantaneous resin polymerizationand vitrification occurs at the surface, whichlowers the polymerization rate and progressivelystops the reaction. Intensive residual stresses maydevelop in the tooth and in the filling material atthis stage, due to the sudden localized volumereduction in material [2–4]. Moreover, since thesematerials do not naturally bond to enamel ordentin-conditioned surfaces, they need theapplication of an adhesive resin layer to controlthe mismatch at the interface between dentalcomposite and tooth structure [5–7].

In order to avoid clinical failure, such as marginalgap formation and bacteria penetration [8], afterlight induced polymerization these dental bondedareas should be protected from the development ofhigh dental composite stress [3,4,9]. Materialrelaxation, which has been recognized by severalauthors [9,10] to occur in resin based restorativematerials, could be used to reduce stress build-upat the bonded interfaces and within the restoration[9–13]. The rate of composite polymerization isthen a critical parameter determining the levels ofstress frozen in the cured resin [10,11]. A methodfor lowering this rate is to use curing light lampswith variable light intensity levels [11]. Never-theless, since we are also interested in adequatepolymerization of deeper layers, it should berealized that the energy is attenuated with depthand thus, conversion can be at risk in these areas.Polymerization shrinkage of dental composites andtheir complete cure therefore, represent twoaspects of the same problem related to theirclinical use [3,4,12–17]. This means that there isthe paradox that all that is needed improve thedegree of conversion can negatively affect thepreservation of the marginal continuity [11,12,15].There are clear indications from hardness measure-ments that the level of polymerization in thedeeper areas of a composite is not of the samemagnitude as at the top of the irradiated sample[18], however, experimental evidence of sucheffects using direct calorimetric measurementshave not yet been presented in literature and willbe investigated in the present study.

It can be hypothesized that a more homogeneousand stress-free restorative condition could be

attained by imposing slower rates of polymerizationthat can favor material relaxation and avoidexcessive stress build-up [3,4,9,13,19]. The aim ofthe present paper is to calorimetrically analyze thekinetics and level of photopolymerization in lightactivated resins as a function of the depth andexposure times, when medium/low energies areused. This information is helpful in determining upto what thickness an adequate polymerization levelis reached in dentin adhesive and composite whenstress relaxation inducing medium/low lightirradiation energies are used.

Materials and methods

In order to evaluate the composite depth shieldingeffect on relative degree and kinetics of polym-erization, all calorimetric experiments were madeby sequentially overlaying and curing three levels ofcomposite and dentin adhesive. A Mettler differen-tial scanning calorimeter (DSC 25) was used tocharacterize the dental composite’s polymeri-zation kinetics when cured by a halogen light sourceplaced at a fixed distance. The halogen light usedfor the experiment was the Demetron (Kerr, USA).This lamp is equipped with a system of opticalfilters and lenses capable of emitting in the blueregion of light. A dental composite Z250 (3M ESPE)was used and photo-cured in the modified Differ-ential Scanning Calorimeter (DSC) operating atconstant temperature in a nitrogen-purged environ-ment. The halogen lamp light tip was fitted into anappropriate guide at the top of the DSC furnacewith the scope of maintaining unvarying irradiationcondition on the samples and reference pan. Thesample and the reference were irradiated at500 mW/cm2 for 60 and 20 s at a distance of 3 cm.Optical properties and efficiencies of curing lightwere measured and monitored during all sets ofexperiments using a calibrated dental radiometercommonly available to clinicians. Intensity outputvalues at the given tip distance and shieldingconditions were measured with the digital dentalradiometer built into the Optilux 501 curing light(Kerr, USA).

Five samples for each series of tests wereprepared and analyzed. Each series of tests wasbased on four steps:

†

In the first step, a 0.3 mm thick, and about 30 mgin weight, resin composite was placed into analuminum pan (without lid), positioned in theDSC furnace maintained at 37 8C and lightirradiated four times at regular time intervals(20 or 60 s irradiation at 100 s intervals).

A. Apicella et al.956

The thermogram of the unshielded resin compo-site was recorded over the entire duration of thetest;

†

In the second step, a 0.3 mm thick fresh resincomposite was placed into a new aluminum pan,overlaid by the previous fully cured sample(acting as a 0.3 mm translucent inlay shield)and polymerized with the same light cureprogram and procedures described before;

†

Figure 1 Heat flux in DSC isothermal light activated

In the third step, an additional 0.3 mm thickfresh composite layer shielded by the 0.6 thickwafer structure obtained in step two was testedusing the same previous light cure program andprocedures.

polymerization of a resin composite. The fourth signal

† was used as a blank subtraction curve.

Finally, the resulting composite wafer (about1 mm thick inlay) was used to overlay a dentinadhesive sample (Excite, Ivoclar) that was poly-merized in this shielded condition using the samelight cure and DSC procedures.

A separate set of experiments was carried out onunshielded adhesive using the procedure describedin step one.

Under normal operating conditions the thermalevents occurring on the sample in a DifferentialScanning Calorimeter (DSC) are compared to thethermal behavior of a reference sample, which isusually an empty aluminum pan. However, since inthe experimental procedures in this study, the DSCcell was recording both heat due to reaction andradiant heat absorbed during the light exposure, itwas necessary to define a protocol to evaluate ablank curve to be used as a correction for theradiant heat contribution. The blank correctioncurve was registered when no additional appreci-able exothermic response was induced by the lightexposure (exhaustion of all resin reactivity). Blankcurves were registered for each single test.

Evaluation of single layer reactionexotherm and blank corrections

For each of the three differently shielded resincomposite and for the dentin adhesive, thecalorimetrically measured values of the heatflows were normalized to the sample weight. Atypical DSC thermogram (full circles) and the blankcorrection (open circles) are shown in Fig. 1 forthe first irradiation of an unshielded compositelight cured at 37 8C. In the isothermal exper-iments, the heat flow due to the polymerizationreaction (dQ/dt)is was evaluated as a function ofthe cure time from the whole DSC signal (wherelight and polymerization energy contributions are

measured) corrected by subtracting the blankcurve registered for each test. The heat ofreaction emitted during the isothermal cure (Qi)was estimated by time integration of the heat fluxcurve measured for each reactive layer ‘i’, wherethe suffix ‘i’ is 1, 2, 3, ad or ads, and refers tounshielded, 0.3 mm shielded and 0.6 mm shieldedcomposite, and unshielded and 1 mm compositeshielded adhesive, respectively.

Since it was observed that for all material tested,the main part of the reaction occurs at the firstexothermic peak while the second and the thirdlight exposures produced small, but still measur-able, residual exotherms that completelyexhausted any residual reactivity of the material;the DSC signal of the fourth exposure of each testwas considered as a blank correction curve.

Evaluation of degree of conversion andreaction rate

DSC measurements were used to determine theadvancement of the reaction assuming thatthe heat developed during polymerization isproportional to the fraction of reactive groupsconsumed.

A relative estimate of the conversion at differ-ent time exposures and shielding conditions wasevaluated, considering the sum of the threemeasurable reaction exotherms as the maximumheat of reaction attainable under the specific lightcure conditions used in the experiments (seeTables 1–3 for the values and statistical analysisof the relative degrees of conversion for the threeirradiations and depth levels). The relative degreeof conversion during the DSC scans, ai(t), was then

Table 1 Statistics on conversion for the four light scans on the first layer of composite (0.0–0.3 mm depths).

First layer 20 s 60 s

Q1exp, J/g (SD) % (SD) Q1exp, J/g (SD) % (SD)

First irradiation 18.2 (0.5) 73.8 (2.1) 23.2 (1.2) 94.1 (5.0)Second irradiation 1.1 (0.2) 4.4 (0.8) 0.8 (0.3) 3.0 (1.1)Third irradiation 0.4 (0.1) 1.5 (0.4) 0.2 (0.1) 0.7 (0.5)Fourth irradiations 0 0 0 0Q1tot (J/g) 19.6 (0.6) 79.6 (2.4) 24.2 (1.3) 97.8 (5.2)

Table 2 Statistics on conversion for the four light scans on the second layer of composite (0.3–0.6 mm depths).

Second layer 20 s 60 s

Q2exp, J/g (SD) % (SD) Q2exp, J/g (SD) % (SD)

First irradiation 18.3 (0.8) 74.2 (3.0) 23.7 (2.2) 96.2 (8.7)Second irradiation 1.0 (0.2) 4.1 (0.9) 0.8 (0.2) 3.0 (0.6)Third irradiation 0.4 (0.1) 1.6 (0.4) 0.2 (0.1) 0.9 (0.4)Fourth irradiations 0 0 0 0Q2tot (J/g) 19.7 (0.9) 79.9 (3.6) 24.72 (2.31) 100 (9.3)

Light shielding effect of overlaying 957

defined as

aiðtÞ ZQiexpðtÞ

Qtot(1)

where Qiexp(t) is the partial heat reaction devel-oped during the single light activated polymeri-zation test of system ‘i’ and Qtot is the maximumattainable total heat of reaction obtained byadding the exothermal contributions of the threeconsecutive light exposures. Furthermore, accord-ing to the experimental results, the maximumattainable heat of reaction for the resin compositeto be used as reference for all tests was defined asthe mean value of the Qitot reached in the threeshielding conditions.

Results and discussion

The blank corrected exotherms at the firstirradiation step for the shielded and unshielded

Table 3 Statistics on conversion for the four light scans o

Third layer 20 s

Q3exp, J/g (SD) % (SD)

First irradiation 18.8 (1.0) 76.2 (4Second irradiation 1.6 (0.4) 6.5 (1Third irradiation 0.5 (0.3) 2.0 (1Fourth irradiations 0 0Q3tot (J/g) 20.9 (1.6) 84.7 (6

conditions are shown in Fig. 2. The three heat fluxesin this figure are diagrammed as a function of thetime and are indicated as first layer (0.3 mm freshsample), second layer (0.3 mm fresh sampleshielded with 0.3 mm cured composite), and thirdlayer (0.3 mm fresh sample shielded with 0.6 mmcured composite). The samples in Fig. 2 were firstthermally equilibrated in the DSC furnace for 20 sand then light irradiated for 20 s. This light energysupply suddenly activated the resin compositeproducing initially an intense exothermic peak,with its maximum located about 20 s from thestart, which is associated with the polymerizationof the resinous composite component. The reactioncontinued even after the interruption of the lightenergy supply and was progressively exhausted inthe successive 60 s. For the samples indicated assecond and third layers the light shielding throughthe composite [17,18] resulted in a reduction of theintensities of the exothermic peaks. The light energyintensities crossing the 0.3, 0.6 and 1 mm composite

n the third layer of composite (0.6–1 mm depths).

60 s

Q3exp, J/g (SD) % (SD)

.0) 24.1 (1.4) 97.8 (5.6)

.6) 0.7 (0.2) 3.0 (0.8)

.2) 0.3 (0.2) 1.1 (0.8)0 0

.4) 25.1 (1.6) 101.7 (6.5)

Figure 2 Blank corrected DSC thermograms. Effect ofthe depth light shielding on polymerization exotherms.First irradiation light scans at three different depths.0.3 mm for each layer added.

Figure 3 Effect of the time irradiation: 20 (lowercurve) and 60 s (upper curve) on the kinetics ofpolymerization at 37 8C.

A. Apicella et al.958

shields, measured with the digital radiometer builtinto the Optilux 501 curing light, were 380, 260 and150 mW/cm2, respectively, and were in all casessufficiently high to activate resin polymerization.

Integration of these blank corrected exothermiccurves gives the overall heat of reaction for theresin composite under the different operatingconditions. Tables 1–3 report the statistics on theheats and levels of polymerization reached as afunction of the light exposure durations andrelative positions (depth shielding effect).

The means of maximum heat of polymerizationmeasured for the unshielded, 0.3 and 0.6 mm lightshielding conditions were 24.1 (1.3), 24.7 (1.3) and25.2 (1.6), respectively. The Anova analysis atpZ0.05 did not show any significant statisticaldifference between the three sets of data, eitherfor the conversion reached at each irradiation ordepth, suggesting that the influence of the increas-ing composite shielding was only in lowering thesystem reaction rates [16,18]. The maximumattainable heat of reaction Qtot [J/g] for 60 sirradiation time was then evaluated as a meanvalue between Q1tot, Q2tot and Q3tot. This value, 24,67 J/g, had been used in Eq. (1) to normalize allcomposite conversion data.

The data reported in the three previous tablesare summarized in Table 4 showing that the

Table 4 Statistics on the different levels of polym-erization reached in the resin composite as a functionof the irradiation time and number of light scans.

20 s 60 s

Qtot (J/g) (SD) 20.1 (1.2) 24.7 (2.0)First irradiation 74.7 (3.2) 96.0 (5.1)Second irradiation 79.7 (4.9) 99.1 (6.2)Third irradiation 81.4 (5.0) 100.0 (7.5)

maximum level of polymerization up to almost1 mm depth is only influenced by the lightirradiation time and not by the depth level atwhich polymerization occurs.

The influence of the irradiation time (20 and 60 s)on the advancement and kinetics of polymerizationfor the dental composite is shown in Fig. 3. After theshorter light exposure (first 20 s), the reactioncontinues, even if not supported by light activation,up to about 40 s at the same rate (slope of the curve)of the sample still irradiated (60 s light cure). Afterthis initial almost coincident part, the reaction rateof the shortly irradiated samples begins to slow downreaching its maximum conversion at about 90% of themaximum attainable value.

Fig. 4 shows the polymerization kinetics ofsamples cured for 20 (a) and 60 s (b) at differentshielding depths (first irradiation). The shieldingeffect at different depths [18] results, for bothirradiation times, in a reduction of the reactionrates (slope of the curves in Fig. 4). The overalllevel of polymerization, however, is not influenced

Figure 4 Polymerization kinetics and relative conver-sions for: (a) 20 and (b) 60 s light exposures at threedepths (0.3 mm increments). Conversions are referred toas the maximum attainable overall heat of reaction.

Figure 5 Conversion rates at the three selected depths.

Table 5 Maximum rates of polymerization at differ-ent polymerization depths.

First layer Second layer Third layer

0.040 [1/s] 0.032 [1/s] 0.029 [1/s]

Light shielding effect of overlaying 959

by the shielding effect and it reaches the same finalvalue for both the 20 s light cured materials and 60 scured materials. The repeated 20 s irradiationconditions, however, let the material exhaust allreactivity reaching about 80% of the maximum thatcan be obtained by a prolonged light cure (60 sirradiation).

The reaction rates (da/dt) of the compositeduring polymerization in the three different lightshielding conditions are shown in Fig. 5. Thesetraces were obtained by analytical derivation of thedegree of polymerization curves reported in Fig. 4b(60 s light cure). A maximum reaction rate for eachlayer can then be evaluated at the maximum of thecurves reported in Fig. 5. This maximum occurs forall three shielding conditions at the initial stages ofthe reaction (between 15 and 18 s after initiation oflight irradiation). The values of the maximumreaction rate are reported in Table 5. The compo-site shielding produced a reduction of 20% of themaximum reaction rate for 0.3 mm thicknessand 27% for 0.6 mm thickness. Fig. 5 also gives

Table 6 Statistics on conversion for the four light scansdentin adhesive.

Adhesive Shielded

Qadexp, J/g (SD) % (SD)

First irradiation 214.3 (4.2) 88.7 (1.7)Second irradiation 16.1 (1.2) 6.7 (0.5)Third irradiation 10.6 (0.6) 4.4 (0.2)Fourth irradiations 0 0Qtot (J/g) 240.9 (5.9) 99.8 (2.4)

an estimate of the relative reaction rates of thedifferently shielded materials over the entire rangeof the experimental test time. The shieldedsamples, even if initially presenting a lower rateof polymerization, are able to sustain the reactionup to the same final conversions at final ratesrelatively higher than those of the unshielded resincomposite. As shown at the right hand side of Fig. 5,the rate of reaction at longer times is higher forpolymerization carried out at deeper levels, wherelower energies are supplied [18] due to the higherconcentration of the residual reactive groups,

The influence of 1 mm composite light-shieldingon the advancement of polymerization for thedentin adhesive used in this study is summarizedin Table 6. The comparison between the cumulativeheat generated by polymerization of the unshielded(full circles) and shielded (open circles) resin in thefirst 60 s irradiation run is illustrated Fig. 6. Thereaction rate for the shielded and unshieldedadhesives are almost the same for about 15 s(about 50% conversion), then the shielded resinstarts to slow down, finally reaching 89% of its finalmaximum value. The cumulative heat of reaction ofthe shielded resin increased to about 96% of thefinal value during the second irradiation. The thirdirradiation of the composite shielded adhesiveleads to complete cure of the resin. The corre-sponding conversions for the unshielded adhesivereached about 96 and 99%, respectively, at the firstand second irradiation.

The influence of light shielding on the reactionrate of the adhesive is different from that of theresin composite; in fact, the adhesive in theshielded and unshielded condition shows an almostsimilar initial reaction rate, which starts to differ-entiate only after 15–18 s of continuous irradiation.The adhesive is characterized by a very high heat ofreaction of 242.0 J/g when compared to that of theresin composite, which is 24.7 J/g. Due to theabsence of inert filler and the high concentration ofreactive sites, it can be hypothesized that, for theunshielded resin adhesive, a very high initialconcentration of polymerization radicals which

on the unshielded and 1 mm thick composite shielded

Unshielded

Qadsexp, J/g (SD) % (SD)

233.7 (2.1) 96.7 (0.9)8.1 (0.1) 3.3 (0.4)1.2 (0.1) 0.5 (0.1)

243.0 (2.4) 100.6 (1.0)

Figure 6 Blank corrected heat fluxes for firstirradiation in a 1 mm thick composite shielded, andunshielded dentin adhesive. Sixty seconds light cure at500 mW/cm2.

A. Apicella et al.960

favor the termination reaction, leads to the samereaction rate as the less light activated shieldedsamples. This balancing condition, however, is lostafter the initial stages of the reaction and the lessenergy irradiated samples show a slow reactionrate, in the remaining part of the light curing.Conversely, the resin composite, which is charac-terized by the presence of the filler that lowers theinitial concentration of polymerization radicals, issensitive to light shielding from the initial stages ofthe cure.

Conclusions

It can be deduced from these experiments on resincomposite and dentin adhesive that the shieldingeffect initially reduces the absorbed energy, redu-cing the concentration of light activated initiatorfree radicals, hence lowering the initial reactionrate. The light intensities crossing the 0.3, 0.6 and1 mm composite shields were in all cases suffi-ciently high to start the polymerization of the resincomposite and dentin adhesive, suggesting that theconcentration of light activated radicals is, none-theless adequate to activate and sustain thepolymerization reaction up to complete conversion.The final degree of conversion remains the same forthe three depths of cure, showing no effect ofshielding on the final level of polymerizationreached. This can be attributed (for thickness upto 1 mm and for the light intensities used) to therate of free radical termination that does notprevail on the propagation mechanism (increase inthe lifetime of these radicals). Due to the higherconcentration of the residual reactive groups, therate of reaction at longer times is higher for thepolymerization carried out at deeper levels and,

hence, at lower energies. This event leads to a morehomogeneous and uniform conversion though thewhole restoration.

For thicknesses greater than 1 mm, it could behypothesized that a still lower number of initiators,insufficient for propagation, participate in thereaction and therefore a lower degree of conversioncan be foreseen.

These considerations on the polymerizationmechanism concur with clinical aims for dentalrestorative materials curing:

†

shorter irradiation time; † high and uniform conversion throughout the

whole restoration;

† low shrinkage stress.

These experiments have shown that for thicknessup to 1 mm, the reduction of the energy transferredto the composite due to the depth shielding effectonly has an effect on the kinetics of polymerizationwithout modifying the overall level of polymeriz-ation. Moreover, shorter irradiations, even if notleading to complete conversion, are able to initiateand sustain polymerization to acceptable level ofcure. According to the previous discussion, slowerpolymerization obtained by using lower poweredlamps or thicker layers of composites, could evenreduce the stresses induced by too-fastpolymerizations.

References

[1] Hickel R, Manhart J. Longevity of restorations in posteriorteeth and reasons for failure. J Adhes Dent 2001;3:45–64.

[2] Lutz F. State of the art of tooth-coloured restoratives. OperDent 1996;21:237–48.

[3] Davidson CL, de Gee AJ, Feilzer A. The competitionbetween the composite–dentin bond strength and thepolymerization contraction stress. J Dent Res 1984;63:1396–9.

[4] Ausiello P, Apicella A, Davidson CL, Rengo S. 3D-finiteelement analyses of cusp movements in a human upperpremolar, restored with adhesive resin-based composites.J Biomech 2001;34:1269–77.

[5] Buonocore M, Wileman W, Brudevold F. A report on a resincomposition capable of bonding to human dentine surfaces.J Dent Res 1956;35:847–51.

[6] Fusayama A, Kono A. Improvement of adhesive dentalmaterials. Shikai Tembo 1985;65:77–80.

[7] Van Meerbeek B, Inokoshi S, Braem M, Lambrechts P,Vanherle G. Morphological aspects of the resin–dentininterdiffusion zone with different dentin adhesive systems.J Dent Res 1992;7:1530–40.

[8] Braem M, Lambrechts P, Van Doren V, Van Herle G. In vivoevaluation of four posterior composites: quantitative wearmeasurements and clinical behavior. Dent Mater 1986;2(3):106–13.

Light shielding effect of overlaying 961

[9] Apicella A, Di Palma L, Aversa R, Apicella D. Compasrison ofdifferent light source in shrinkage stress phenomena. J AdvRes 2003;35:13–18.

[10] Uno S, Asmussen E. Marginal adaptation of a restorativeresin polymerized at reduce rate. Scand J Dent Res 1991;October 99:440–4.

[11] Grace G, Mori G, Casa de’Martinis L. Curing light intensityand marginal leakage of resin composite restorations. QuintInt 1996;27:355–62.

[12] Braem M, Lambrechts P, Vanherle G. Clinical evaluation ofdental adhesive systems, part II: a scanning electronmicroscopy study. J Prosthet Dent 1986;55:551–60.

[13] Davidson CL, de Gee AJ. Relaxation of polymerizationcontraction stresses by flow in dental composites. J DentRes 1984;63:146–9.

[14] Feilzer AJ, De Gee AJ, Davidson CL. Relaxation ofcontraction shear stress by hygroscopic expansion. J DentRes 1990;69:36–9.

[15] Ruggeberg F. Contemporary issues in photocuring. CompendContin Educ Dent Suppl 1999;S4–S15.

[16] Harris JS, Jacobsen PH, O’Doherty DM. The effect of curinglight intensity and test temperature on the dynamicmechanical properties of two polymer composites. J OralRehabil 1999;26:635–9.

[17] Halvorson RH, Erickson RL, Davidson CL. An energyconversion relationship predictive of conversion profilesand depth of cure for resin-based composite. Oper Dent2003;28:307–14.

[18] Davidson-Kaban S, Davidson C, Feiltzer A, de Gee A. Theeffect of curing light variation on bulk curing and wall-towall quality of two types and various shades of resincomposites. Dent Mater 1997;13(6):626–31.

[19] Ausiello P, Apicella A, Davidson CL. Effect of adhesive layerproperties on stress distribution in composite restorations—a 3D finite element analysis. Dent Mater 2002;18(4):295–303.