d e n t a l m a t e r i a l s 2 3 ( 2 0 0 7 ) 218–225

avai lab le at www.sc iencedi rec t .com

journa l homepage: www. int l .e lsev ierhea l th .com/ journa ls /dema

Microleakage of porcelain veneer restorations bonded toenamel and dentin with a new self-adhesive resin-baseddental cement

Gabriela Ibarraa,∗, Glen H. Johnsona, Werner Geurtsena, Marcos A. Vargasb

a School of Dentistry, Department of Restorative Dentistry, University of Washington, Seattle, WA, USAb College of Dentistry, Department of Operative Dentistry, University of Iowa, Iowa City, USA

a r t i c l e i n f o

Article history:

Received 29 September 2005

Received in revised form

19 December 2005

Accepted 10 January 2006

Keywords:

Self-adhesive resin cement

Ceramic veneers

Microleakage

Dentin adhesion

Enamel adhesion

a b s t r a c t

Cementation technique of bonded ceramic restorations is a time-consuming and technique-

sensitive procedure critical to long-term success.

Objective. Evaluate the performance of a self-adhesive, modified-resin dental cement (Rely-

X UniCem, 3M-ESPE) for the cementation of ceramic veneer restorations without previous

conditioning of the tooth surface, and in combination with a one-bottle adhesive and a

self-etching adhesive.

Methods. Thirty-six premolars received a veneer preparation that extended into dentin.

Leucite-reinforced pressed glass ceramic (Empress 1) veneers were cemented follow-

ing manufacturers’ instructions, according to the following treatment groups (n = 9): (1)

Variolink–Excite Ivoclar–Vivadent (V + E control), (2) Unicem + Single Bond 3M-ESPE (U + SB),

(3) Unicem + Adper Prompt L-Pop 3M-ESPE (U + AP), (4) Unicem 3M-ESPE (U). After 24 h stor-

age at 37 ◦C, teeth were thermocycled (2000 cycles) at 5 and 55 ◦C, immersed in ammoniacal

silver nitrate for 24 h, placed in a developer solution overnight and sectioned using a slow-

speed saw. Three 1 mm longitudinal sections were obtained from each tooth and evaluated

for leakage with a microscope (1× to 4×). Imaging software was used to measure stain pen-

etration along the dentin and enamel surfaces.

Results. ANOVA with SNK (˛ = 0.05) revealed that on dentin, U had significantly less leakage

than U + SB and U + AP, but no different than V + E; on enamel U had leakage values that were

significantly greater than the groups with adhesives.

Significance. The self-adhesive cement U gave low leakage on dentin that was comparable

to the cement that employed an adhesive for sealing dentin, whereas this cement benefits

from use of an adhesive when cementing to enamel.

© 2006 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

1. Introduction

The use of bonded ceramic restorations in dentistry hasincreased appreciably due to the development of adhesive

∗ Corresponding author at: Department of Restorative Dentistry, UniveWA 98195-7456, USA. Tel.: +1 206 543 5948; fax: +1 206 543 7783.

E-mail address: gibarra@u.washington.edu (G. Ibarra).0109-5641/$ – see front matter © 2006 Academy of Dental Materials. Pudoi:10.1016/j.dental.2006.01.013

rsity of Washington, 1959 N.E. Pacific Street, Box 357456, Seattle,

materials that allow for more conservative restorative tech-niques as well as the ability of achieving excellent estheticappearance and adequate strength [1]. Among these bondedceramic restorations, ceramic veneers have gained popularity

blished by Elsevier Ltd. All rights reserved.

2 3

aio

astcrdieIcahm

Utdomtsipop

bav

w[im

taao

emo[btppdmr

tacc

(750 mW/cm2).

d e n t a l m a t e r i a l s

s a conservative restoration, where a thin ceramic coverings bonded, preferentially to enamel, after minimal preparationn tooth structure [2,3].

The cementation technique, which is a time-consumingnd technique-sensitive procedure, is key to the long-termuccess of these types of restorations. The strength andhe durability of the bond between the porcelain, the lutingement and the enamel/dentin interface play an importantole in the outcome of ceramic veneers, particularly whenentin is involved [3]. It is not uncommon that, particularly

n the gingival third of a veneer preparation, dentin will bexposed due to the thin layer of enamel present at this site [4].n this case, the cementation procedure becomes even moreritical because high failure rates in veneers have been associ-ted to large exposed dentin surfaces and the cervical marginas been regarded as a problematic area to achieve perfectarginal adaptation [3].A self-adhesive, resin-based dental cement (Rely-X

niCem, 3M-ESPE), which advocates no pre-treatment ofooth surfaces, thus simplifying the cementation proce-ure, has recently been introduced. This cement has anrganic matrix composed of multi-functional phosphoric acidethacrylates, which react with inorganic fillers (72 wt.%)

hat are basic in nature or with hydroxyapatite from toothtructure. Water that is released from the setting reactions thought to play a role in its neutralization, raising theH value from 1 to 6. The setting of the cement is basedn a free radical polymerization reaction initiated by eitherhotoactivation or a redox system [5,6].

The cement has been recommended for luting all metal-ased and ceramic crowns, as well as partial coverage ceramicnd indirect composite restorations, with the exception ofeneers [5,6].

Good marginal adaptation of all-ceramic crowns cementedith Rely-X UniCem to dentin has already been documented

5]. Preliminary studies have shown good results when bond-ng pressed ceramic inlay restorations to dentin and enamel

argins [7].If the self-adhesive cement could be used predictably for

he cementation of ceramic veneer restorations, it would serves a user-friendly universal cement. However, more studiesre needed before final recommendations for the clinical usef this cement can be made.

The clinical success of cemented restorations has beenvaluated by measuring marginal fit and microleakage forany years, in spite of the fact that there is no restoration

r luting material able to achieve a complete marginal seal8,9]. In the case of all-ceramic restorations, microleakage haseen correlated with the loss of the integrity of the bondo tooth structure, and this has been associated with otherroblems such as secondary caries, post-operative sensitivity,ulpal inflammation, staining and plaque accumulation [9–11]ue to the clinically undetectable passage of bacteria, fluids,olecules or ions between tooth structure and the cemented

estoration [12].The aim of this study was to test the hypothesis that

he application of a new self-adhesive resin cement, useds a luting agent, would result in good marginal integrity oferamic veneers to dentin as well as enamel, without prioronditioning of the tooth surface or in combination with

( 2 0 0 7 ) 218–225 219

other adhesive systems that previously condition enamel anddentin.

Altogether, the hypothesis tested was that the microleak-age of the new self-adhesive luting cement is similar to aconventional resin cement when bonding a porcelain veneerto enamel and dentin.

2. Materials and methods

Thirty-six human premolars, previously stored in aNaN3 + NaCl solution for no more than 6 months, wereprepared for porcelain veneer restorations. The preparationswere made with a #834 016 bur (Brasseler, Savannah, GA31419, USA) to establish a 0.3 mm depth cut, and finishedwith a fine rounded tip diamond #6844 016 (Brasseler, Savan-nah, GA 31419, USA). The preparation’s margins ended asbutt joint at the incisal edge and a chamfer that extendedapproximately 1 mm beyond the CEJ. An impression ofeach of the preparations was made with a vinyl polysiloxaneimpression material (Aquasil LV and Aquasil, Caulk, Dentsply;Lot 020608) using copper rings for material retention. Theteeth were then stored in artificial saliva solution and theimpressions were sent to a commercial laboratory (Nakan-ishi Dental Laboratory Inc., Seattle, WA). Leucite-reinforcedpressed ceramic (Empress I) porcelain veneer restorationswere fabricated in an A-1 shade based on the Vita shadeguide, etched with hydrofluoric acid and silanated in thelaboratory.

The teeth were then divided into four treatment groupsaccording to the cementation procedure (Table 1). Group1: restorations cemented using a conventional resin-basedcement and its proprietary adhesive system as a con-trol (Variolink II [Base Yellow 210/A3, Lot E 43489; Cata-lyst Lot E 34696] and Excite [Lot E41824], Ivoclar/Vivadent).Group 2: restorations cemented using self-adhesive, modified-resin dental cement (Rely-X UniCem Maxicaps, 3M-ESPE;Lot 143650) in combination with enamel and dentin con-ditioning with phosphoric acid and a one-bottle adhe-sive system (Single Bond, 3M-ESPE). Group 3: restorationscemented using a self-adhesive resin-based dental cement(Rely-X UniCem Maxicaps, 3M-ESPE; Lot 143650) in combi-nation with a self-etching adhesive system (Adper PromptL-Pop, 3M-ESPE; Lot 147563) to condition enamel and dentin.Group 4: restorations cemented with a self-adhesive resin-based dental cement (Rely-X UniCem Maxicaps, 3M-ESPE;Lot 143650) without previous conditioning of enamel anddentin.

Before cementation, all veneers were tried-in, cleaned withphosphoric acid and re-silanated (Rely-X Ceramic Primer; Lot2721) in accordance with clinical practice.

Light-curing of the adhesives and the cement was carriedout with a Demetron 401 light unit (Demetron/Kerr, Dan-bury, CT) as indicated in Table 1. Light output was mea-sured every six samples to ensure proper resin polymerization

The specimens were stored in artificial saliva at 37 ◦C for72 h and were thermocycled in 5 and 55 ◦C water temperaturesfor 2000 cycles with 20 s dwell time at each temperature and atransfer time of 10 s, for a total of 60 s per cycle.

220 d e n t a l m a t e r i a l s 2 3 ( 2 0 0 7 ) 218–225

Table 1 – Application technique for seating ceramic veneer restorations with different adhesive systems and lutingcements

Treatment Materials

Excite/Variolink Single bond/Unicem Prompt L-pop/Unicem Unicem

Veneer preparation Etch with H3PO4 × 60 s Etch with H3PO4 × 60 s Etch with H3PO4 × 60 s Etch with H3PO4 × 60 sRinse × 30 s Rinse × 30 s Rinse × 30 s Rinse × 30 sAir dry Air dry Air dry Air drySilanate internalaspect × 60 s

Silanate internalaspect × 60 s

Silanate internalaspect × 60 s

Silanate internalaspect × 60 s

Allow solvent toevaporate

Allow solvent toevaporate

Allow solvent toevaporate

Allow solvent toevaporate

Apply adhesive tointernal aspect

Apply one coat ofadhesive to internalaspect

Lightly scrub one layerof adhesive to internalaspect: 15 s

Air-thin Air-dry for 2–5 s Air-thinDO NOT light cure DO NOT light cure DO NOT light cure

Substrate preparation H3PO4: etchenamel × 15–30 s anddentin × 10–15 s

H3PO4: etch enamel anddentin × 10 s

Slightly dry surface Leave dentin andenamel surfacemoist/glossy

Rinse: 5 s at least Rinse: 30 s Adhesive: lightly scrubone layer on enameland dentin: 15 s

Remove excesswater—do not over-drydentin

Blot excess water—leavemoist surface

Air-thin

Adhesive: apply severallayers

Adhesive: apply 2 coats Adhesive: apply asecond coat withoutrubbing

Air-dry: 1–3 s. Avoidpooling

Air dry: 2–5 s Air dry

Cure: 20 s DO NOT light cure DO NOT light cureSeating instructions Apply base cement to

veneer and toothApply mixed cement toveneer

Apply mixed cement toveneer

Apply mixed cement toveneer

Maintain pressure forseveral seconds andtack: 10–20 s

Tack-cure excesscement × 2–4 s andremove

Tack-cure excesscement × 2–4 s andremove

Tack-cure excesscement × 2–4 s andremove

Remove excess cement Light-cure × 40 s Light-cure × 40 s Light-cure × 40 s

Fig. 1 – Section of the crown showing the interface of

with brushLight-cure × 40 s

2.1. Microleakage evaluation

The apices of the roots were sealed with an acrylic resin(Duralay Inlay Pattern Resin, Reliance) and the teeth were thencoated with two layers of quick dry nail varnish that extendedup to 1 mm from the margins of the ceramic veneer restora-tions. Care was taken not to over dry the enamel surroundingthe margins while the nail varnish dried.

The teeth were placed in 50 wt.% ammoniacal silvernitrate for 24 h, rinsed extensively with water, and placed infreshly mixed developer solution (Kodak Developer D-76, CAT1464817, 0251 C5 02749) under a strong light for 12 h. Afterrinsing them with water and sand blasting the porcelain sur-face carefully, a 1 mm layer of composite was bonded to theveneer (All Bond 2, Bisco, Lot 0200002521; Filtek Z 250 B 1 shade,3M-ESPE, Lot 9BB) to provide sufficient bulk for handling, andlight-cured for 40 s. The roots were removed and the crownswere sectioned in a cervical–incisal direction with a diamond

blade to obtain three slices (∼1 mm) from each tooth (Fig. 1).

Sections were analyzed for leakage at the cervical andincisal margins by means of a light microscope (Nikon EclipseE400, Japan) at 1×, 2× and 4×, using an image analysis com-

enamel (E) and dentin (D) with the porcelain veneer (PV)and the overlying composite (C). Penetration of theammoniacal silver nitrate can be observed on the dentinside (arrow).

2 3 ( 2 0 0 7 ) 218–225 221

ptsdt

2

S(fsIahtott

2

AfacocBtasasuH

3

3

Aofbos

iBU(athns

t

Fig. 2 – Mean percent microleakage values betweenporcelain veneers and dentin. The numbers identify mean

The samples treated only with unicem (U) showed a gapbetween the cement and the enamel, which is in accor-dance with the high leakage values observed in this group(Figs. 7 and 8).

d e n t a l m a t e r i a l s

uter program (Meta Vue, Universal Imaging Corp., Downing-own, PA). Microleakage values were obtained by measuringtain penetration for the total surface length, separately forentin and enamel, and were expressed as a percentage ofhe total length of the veneer preparation.

.2. Statistical analysis

ince there were four different cement-bonding combinationsi.e., treatments) and samples were independent, a single-actor analysis of variance model was employed to test forignificant main effects, separately for dentin and for enamel.f main effects were significant (˛ = 0.05) and test for equal vari-nce was not significant, the Student–Newman–Kuel’s postoc test for multiple comparison of means was conductedo determine which means differed. When the assumptionf equal variances was not met, the Games–Howell post hocest was used to identify which means differed. All hypothesisesting was conducted at the 95% level of confidence.

.3. SEM evaluation

fter cementation, one tooth from each group was preparedor SEM evaluation. The teeth were stored in water for 4 weeksnd were fixed for 72 h in 2.5% glutaraldehyde in 0.1 M Na-acodylate buffer. Cross-sections of an approximate thicknessf 1 mm (Fig. 1) were obtained from the teeth using a water-ooled slow-speed diamond saw (Buehler Isomet 1000TM,uehler Ltd., Lake Bluff, IL, USA) and each section was sequen-ially polished with 600 and 800 grit of silicon carbide paper, 6nd 1 �m diamond slurries and 0.04 �m aluminum oxide. Thepecimens were dehydrated in an ascending series of ethanolnd critical-point dried with HMDS, mounted on aluminumtubs and gold-sputter coated to prepare them for analysisnder a field-emission scanning electron microscope (FE-SEM,itachi S-4000).

. Results

.1. Microleakage analysis

total of 144 specimen sections were available for evaluationf microleakage at the interface. Three sections were obtainedrom each of the 48 teeth and information was gathered fromoth sides of each section. Microleakage was observed in mostf the specimens, especially on the dentin side, which is con-istent with existing evidence [13–18].

Mean microleakage values for the dentin side were 44.1%n the Variolink and Excite group (E + V), 55.5% in the Singleond and Unicem group (SB + U), 54.7% in the Prompt andnicem group (AP + U) and 28.1% in the Unicem group (U)

Fig. 2). ANOVA test was significant and the test for equal vari-nces was not significant. The Student–Newman–Kuel’s (SNK)est to compare means (˛ = 0.05) revealed that on dentin: Uad significantly less leakage than SB + U and AP + U but was

ot different than E + V and that E + V, SB + U, AP + U were nothown to differ.

The microleakage values for the enamel side were 2.5% inhe E + V group, 3.1% in SB + U group, 2.2% in the AP + U group

subsets not shown to differ at ˛ = 0.05. The vertical barsshow the value of a single standard deviation.

and 10.8% in the U group (Fig. 3). ANOVA test for main effectswas again significant, as was the test for equal variancesdue to a higher standard deviation associated with enamelmicroleakage for U. For this reason, the Games–Howell test(˛ = 0.05) was employed to compare means. This test revealedthat on enamel, U had leakage values that were significantlygreater than any of the groups where an adhesive systemwas used in combination with the luting cement. In the lattergroups, leakage was minimal and no statistically significantdifference was found amongst them (Fig. 3).

3.2. Scanning electron microscopy analysis

The interface of the samples cemented with E + V, SB + Uand AP + U showed good adaptation of the cement to theenamel surface. No gap formation between the cement andthe enamel was evident in the E + V group, which is consistentwith the microleakage values obtained for these specimens(Fig. 4). It appeared that the use of an adhesive resulted inconsistently good adaptation of the cement to the enamel,regardless of the type of conditioning, as was demonstratedby similar leakage values when using phosphoric acid and theself-etching adhesive system (Figs. 5 and 6).

Fig. 3 – Mean percent microleakage values betweenporcelain veneers and enamel. The numbers identify meansubsets not shown to differ at ˛ = 0.05. The vertical barsshow the value of a single standard deviation.

222 d e n t a l m a t e r i a l s 2 3 ( 2 0 0 7 ) 218–225

Fig. 4 – Scanning electron microscopy image showing goodadaptation at the enamel–cement interface (arrows) in asample cemented with E + V. An almost imperceptibletransition seemed to take place at theenamel–cement–porcelain interface, where minimal

Fig. 6 – Scanning electron microscopy image showing goodadaptation at the enamel–cement interface (arrows) in asample cemented with AP + U. Minimal leakage valueswere also observed in these set of specimens.

leakage values were observed.

Fig. 5 – Scanning electron microscopy image showing goodadaptation at the enamel–cement interface (arrows) in asample cemented with SB + U. An almost imperceptibletransition seemed to take place at theenamel–cement–porcelain interface, where minimalleakage values were also observed.

Fig. 7 – Scanning electron microscopy image showing gapformation at the enamel–cement interface (arrows). Animperceptible transition seemed to take place at the

cement–porcelain interface, where no leakage wasobserved.

4. Discussion

The use of dyes is one of the oldest techniques to measuremicroleakage and the use of a 50 wt.% silver nitrate solutionhas been considered an acceptable technique for this purpose

d e n t a l m a t e r i a l s 2 3

Fig. 8 – Scanning electron microscopy image showing gapformation at the enamel–unicem interface (arrows). Animperceptible transition seemed to take place at thecement–porcelain interface, where no leakage waso

[piottciaa∼

4

WUtcocepc

eAp(auw

bserved.

19]. A disadvantage of this tracer is that the silver nitratearticle is an extremely small particle that measures approx-

mately 0.059 nm in radius and the solution has an acidic pHf ∼4.2 [9,20]. Therefore, penetration of the silver particle athe interface is frequently observed and it has been suggestedhat it may be greater because of dissolution of remnant cal-ium phosphate salts at the adhesive interface, resulting inncreased porosity due to a light etching effect by the mildlycidic solution. To avoid this potential drawback, the use ofbuffered solution of ammoniacal silver nitrate with a pH of9.5, has been reported [20] and was used in the current study.

.1. Dentin interface

ithout any conditioning, the self-etching cement Rely-Xnicem (3M-ESPE) (U) showed improved sealing of dentin at

he cervical margin when compared to a conventional resinement for which the smear layer was removed by the usef phosphoric acid, although this was not statistically signifi-ant in this study. These findings are in accordance with Behrt al. [5] who found similar marginal adaptation based on dyeenetration and SEM replica analysis, to that obtained withonventional cements on dentin margins.

The use of two different adhesive systems, a one-step total-tch (Single Bond, 3M-ESPE) and a one-step self-etch (Rely-Xdper Prompt, 3M-ESPE), to condition the enamel and dentinrior to cementation, did not improve the sealing ability of

U) in dentin, when compared to the control. De Munck etl. [21] reported similar �TBS values when bonding to dentinsing Rely-X Unicem without previously etching the dentinith H3PO4 or a conventional cement as a control. However,

( 2 0 0 7 ) 218–225 223

when the dentin was etched prior to cementation with Rely-X Unicem, the �TBS values significantly decreased. Similarly,in the present study, when the dentin was pre-treated witheither H3PO4 or an acidic monomer from the self-etching sys-tem, increased leakage was observed. Pre-etching may removeall of the buffer capacity of dentin, interfering with its abilityto raise the pH of the acidic resin as it sets, thereby loweringits conversion (David Pashley, personal communication).

According to Behr et al. [5], in images obtained with trans-mission electron microscopy, a lack of a hybrid layer is evidentat the dentin interface when the self-adhesive cement is used.This agrees with De Munck et al. [21] who reported no evidenceof dentin demineralization even considering the initial low pHof the cement (pH <2, according to the manufacturer). Thisresult in the absence of a hybrid layer which, although thinner,is present when self-etching adhesive systems are used [22].Indeed, if the acid resin cement does not penetrate throughthe smear layer, the strength of the bond may be limited tothe strength of the smear layer.

In this study, the cement was never used alone after etch-ing dentin, but in combination with either a simplified total-etch single-bottle adhesive or a one-step self-etching adhe-sive, which coated the pre-treated dentin with a resin layer. Inboth cases, there were increased leakage values over thoseseen when the self-adhesive cement was used alone. It isknown that the previously reported incompatibility of sim-plified adhesives, especially single-step self-etch adhesives,and dual/auto-cured luting cements may be due, in part, toadverse acid–base reactions between the acidic monomersand the basic tertiary amines that are commonly used as cata-lysts in these systems [23]. The tertiary amines are consumedas they come in contact with the acidic resin monomersdue to the absence of an additional adhesive resin layer,which decreases their capacity to generate free radicals forthe polymerization reaction. To solve this problem, alterna-tive redox catalyst systems have been introduced, such asthe commonly used sodium salt of aryl sulphinic acid [24].Many single-bottle adhesive systems have an additional bot-tle of an activator solution that contains such a salt. Theone-step self-etch adhesives are even more susceptible tothis adverse chemical reaction because they are inherentlymore acidic due to a higher content of acidic resin monomers[24].

On the other hand, the one-bottle or simplified adhesives,either total-etch or self-etch, behave as permeable mem-branes due to the absence of a more hydrophobic resin coat,such as is used in conventional systems [25]. Transudationof dentinal fluid across these simplified adhesives has beenobserved, and it has been speculated that this may interferewith the proper polymerization of the luting resin cement byinducing emulsion polymerization of the coupling hydropho-bic resin cements [23,25]. It has also been reported that theslow-setting of the dual/auto-cured luting cements may allowsufficient movement of dentinal fluid to cause microblistersat the resin–composite interface, which may result in subse-quent failure of the bond because these blisters will eventually

act as stress raisers [25]. As a result of these mechanisms, alack of a hermetic seal at the tooth structure–adhesive inter-face will probably occur. In the current study, the seal at thecervical margin of the cemented veneer restorations when

l s 2

r

224 d e n t a l m a t e r i a

using the adhesive systems was less effective than when thecement was used alone.

Furthermore, it has been reported that the one-step self-etching adhesive system (AP) has shown better sealing whensubsequent coats of adhesive are applied. Light-curing thefirst coat on dentin before applying the second coat, has beenrecommended to assure adequate resin polymerization [26].Nevertheless, Tay et al. [25] more recently reported that denti-nal fluid transudation is still observed after additional layersof simplified adhesive systems are applied, even if they arelight-cured separately. In the present study, the one-step self-etching adhesive (AP) was used following the manufacturer’sinstructions (Table 1) which now recommend application oftwo consecutive coats of adhesive without light-curing thefirst coat. Interestingly, there was no statistically significantdifference regarding leakage when compared to the conven-tional system but there was a difference when compared to(U) alone. Leakage was observed, in general, between the toothstructure and the cement. We speculate that there was a defi-cient seal probably due to a lack of complete polymerizationof the luting cement because of the presence of water at theinterface.

When the self-adhesive cement (U) was used on its own,the microleakage values were no different than the valuesobtained with the conventional cement. According to theinformation available on this self-adhesive cement, neutral-ization of the acidic reaction takes place as polymerizationprogresses (ESPE-information from the manufacturer). It isimportant to mention that the small amount of leakage thatwas observed in dentin may not be indicative of adequatelong-term performance of the cement. The durability that self-etching adhesive systems have on cement-dentin bonds aftersome time, remains to be evaluated.

4.2. Enamel interface

Without any conditioning (U) showed significantly greaterleakage at the enamel interface, which was not seen when thecement was used in combination with other adhesive systemsor a conventional cement. This may suggest an insufficientetching ability of the cement to smear layer covered enamel,and therefore, the lack of development of adequate microme-chanical retention. The use of the conventional cement as wellas the one-bottle and the self-etching adhesive systems werelikely to result in good micromechanical retention, since theformer used phosphoric acid prior to the adhesive applicationand the latter was a strong self-etching adhesive (AP) whichhas been shown to adequately etch enamel [27]. These resultsare in agreement with DeMunk et al. [21], who reported lower�TBS to enamel that was not acid-etched previous to the useof Rely-X Unicem, with mostly adhesive failures in the sam-ples, which is probably due to the lack of etching throughthe smear layer into the underlying enamel. In this samestudy, the �TBS values increased significantly when enamelwas previously etched with phosphoric acid, which obvi-ously resulted in the formation of adequate micromechanical

retention.

The role of chemical bonding of the self-adhesive cementwith enamel may be insufficient to obtain an adequate sealbetween the cement and enamel. This is evident in the SEM

3 ( 2 0 0 7 ) 218–225

images, where previously treated enamel by either a phos-phoric acid conditioner or a strong self-etching adhesive sys-tem, show no evidence of gap formation between the cementand the enamel (Figs. 4–6). However, when the self-adhesivecement was used on its own, an evident gap can be observed(Figs. 7 and 8), which may be due to a combination of fac-tors such as inadequate etching through the enamel smearlayer for micromechanical retention on the enamel surface,and the weak cohesive strength of enamel smear layers [28] ofthe cement.

The inadequate formation of micromechanical retentionon enamel may be due to the high viscosity that the cementhas after mixing and the short interaction time that it has withthe tooth surface before light-curing takes place. The initiallow pH (<2) may not be sufficient to etch the enamel if etchingtime is not adequate, and if neutralization reactions take placerapidly.

Another explanation for the presence of gaps at theenamel interface could be lack of adequate pressure duringthe cementation procedure. De Munk [21] has reported animproved adaptation of the cement to the substrate when itis applied under pressure due to its thixotropic characteris-tics. In this study, however, the veneers were cemented underpressure that simulated that applied in the clinical setting, andalthough gap formation was not seen at the dentin interface,it was still evident at the enamel interface.

5. Conclusions

The seal of the self-adhesive, resin-based cement is compa-rable to cements that employ adhesives for sealing dentin,whereas this cement appears to benefit from the use of a con-ventional conditioner, such as phosphoric acid, or a strongself-etching adhesive system when cementing to enamel.

Due to the excessive enamel microleakage observed inthis study when the self-adhesive cement was used alone,the authors would not recommend its use for cementationof ceramic veneers. The hypothesis tested was rejected forenamel substrates.

The lack of adequate performance of this self-etchingadhesive cement on enamel suggests the need to moreextensively investigate the adhesion mechanism of the self-adhesive cement, surface characterization of the substrateand evaluation of its long-term clinical performance.

Acknowledgements

“This investigation was supported in part by 3M-ESPE”.The authors would like to express their gratitude to OHSU

for the thermocycling of the samples and to Dr. David Pashleyfor his valuable contribution.

e f e r e n c e s

[1] Rosenstiel SF, Land MF, Crispin BJ. Dental luting agents: areview of the current literature. J Prosthet Dent1998;80(3):280–301.

2 3

d e n t a l m a t e r i a l s

[2] Peumans M, Van Meerbeek B, Yoshida Y, Lambrechts P,Vanherle G. Porcelain veneers bonded to tooth structure: anultramorphological FE-SEM examination of the adhesiveinterface. Dent Mater 1999;15:105–19.

[3] Peumans M, Van Meerbeek B, Lambrechts P, Vanherle G.Porcelain veneers: a review of the literature. J Dent2000;28:163–77.

[4] Ferrari M, Patroni S, Balleri P. Measurement of enamelthickness in relation to reduction for etched laminateveneers. Int J Periodontics Restorative Dent1992;12(5):407–13.

[5] Behr M, Rosentritt M, Regnet T, Lang R, Handel G. Marginaladaptation in dentin of a self-adhesive universal resincement compared with well-tried systems. Dent Mater2004;20:191–7.

[6] Technical data sheet, 3M-ESPE.[7] Rosentritt M, Behr M, Lang R, Handel G. Influence of cement

type on the marginal adaptation of all ceramic MOD inlays.Dent Mater 2004;20:463–9.

[8] Gladys S, Van Meerbeek B, Lambrechts P, Vanherle G.Microleakage of adhesive restorative materials. Am JDentistry 2001;14(3):170–6.

[9] Alani AH, Toh CG. Detection of microleakage around dentalrestorations: a review. Oper Dent 1997;22(4):173–85.

[10] Hekimoglu C, Anil N, Yalcin E. A microleakage study ofceramic laminate veneers by autoradiography: effect ofincisal edge preparation. J Oral Rehabil 2004;31:265–70.

[11] Cox CF, Felton D, Bergenholtz G. Histopathological responseof infected cavities treated with Gluma and Scotchbonddentin bonding agents. Am J Dent 1988;1:189–94.

[12] Kidd EAM. Microleakage: a review. J Dent 1976;4:199–206.[13] Kanka J. Microleakage of five dentin bonding systems. Dent

Mater 1989;5:415–6.[14] Lacy AM, Wada C, Du W, Watanabe L. In vitro microleakage

at the gingival margin of porcelain and resin veneers. J

Prosthet Dent 1991;67(1):7–10.

[15] Zaimoglu A, Karaagaclioglu L, Uctasli S. Influence ofporcelain materials and composite luting resin onmicroleakage of porcelain veneers. J Oral Rehabil1992;19:319.

( 2 0 0 7 ) 218–225 225

[16] Retief DH. Do adhesives prevent microleakage? Int Dent J1994;44:19–26.

[17] Sim C, Neo J, Kiam Chua EK, Tan BY. The effect of dentinbonding agents on the microleakage of porcelain veneers.Dent Mater 1994;10(4):278–81.

[18] Tay FR, Gwinnett AJ, Pang KM, Wei SHV. Variability inmicroleakage observed in a total-etch wet-bondingtechnique under different handling conditions. J Dent Res1995;74:1168–78.

[19] Hammesfahr PD, Huang CT, Shaffer SE. Microleakage andbond strength of resin restorations with various bondingagents. Dent Mater 1987;3:194–9.

[20] Tay FR, Pashley DH. Water treeing—a potential mechanismfor degradation of dentin adhesives. Am J Dent2003;16(1):6–12.

[21] De Munk J, Vargas MA, Van Landuyt K, Hikita K, LambrechtsP, Van Meerbeek B. Bonding of an auto-adhesive lutingmaterial to enamel and dentin. Dent Mater 2004;20:963–71.

[22] Tay FR, Sano H, Carvalho R, Pashley EL, Pashley DH. Anultrastructural study of the influence of acidity ofself-etching primers and smear layer thickness on bondingto intact dentin. J Adhes Dent 2000;2(2):83–98.

[23] Tay FR, Pashley DH, Suh BI, Carvalho RM, Itthagarun A.Single-step adhesives are permeable membranes. J Dent2002;30:371–82.

[24] Tay FR, Pashley DH, Peters MC. Adhesive permeabilityaffects composite coupling to dentin treated with a self-etchadhesive. Oper Dent 2003;28(5):610–21.

[25] Tay FR, Frankenberger R, Krejci I, Bouillaguet S, Pashley DH,Carvalho RM, Lai CNS. Single-bottle adhesives behave aspermeable membranes after polymerization. I. In vivoevidence. J Dent 2004;32(8):611–21.

[26] Pashley EL, Agee KA, Pashley DH, Tay FR. Effects of oneversus two applications of an unfilled, all-in-one adhesiveon dentin bonding. J Dent 2002;30:83–90.

[27] Pashley DH, Tay FR. Aggressiveness of contemporary

self-etching adhesives. Part II: etching effects on ungroundenamel. Dent Mater 2001;17:430–44.

[28] Tao L, Pashley DH, Boyd L. Effect of different types of smearlayers on dentin and enamel shear bond strengths. DentMater 1988;4:208–16.