Effect of Different Conditioning Protocolson Adhesion of a GiC to Dentin

Franklin R. TayVRJ. Smales^H. NgoVS.H.Y. Wei ^/David H. Pashley^

Purpose: This study examined the ultrastructure and microtenslie bond strength (ÍJTBS) of a restorativeglass-ionomer cement (GIC; Dentsply) to sound dentin that was conditioned with various techniques.

Materials and Methods: Dentin surfaces from extracted human third molars were abraded with 180-gritSiC paper. Five groups of three teetii each were prepared: C - no acid pretreatment (control]; P - 10% poly-acrylic acid IPAA) for 10 s, no rinsing; R - 10% PAA for 20 s, rinsed; K - 25% PAA for 25 s, rinsed; and H -32% phosphoric acid for 15 s, rinsed. TEM was performed on a bonded specimen from each group, usingunstained, undemineralized sections. GIC buildups were made on the remaining teetti, and after storageat 100% humidity for 24 h, the teeth were sectioned for IJTBS and SEM evaluation.

Results: TEM revealed the presence of a structure known as the intermediate layer in all groups. This layercontains metallic salts contributed by both the GIC and dentin. In group C, this layer was restricted to thesmear layer. In groups P and R, intermediate layers could be found above partially demineralized zoneswithin the intertubular dentin. In groups conditioned with mere aggressive protocols ¡K and H), the inter-mediate layers shifted downward to reside within the superficial portions of completely demineralized col-lagen. Group C had statistically lower MTBS (p < 0.05), whiie the other groups were not significantlydifferent from each other. SEM revealed adhesive failures along the dentin surface in group C and mixedfailures in the other groups.

Conclusion: The lower MTBS observed In the control group reflects the weakness of the smear layer at-tachment to dentin. The higher pTBS in the other groups probably represent the cohesive strength of GICunder tension, rather than its true adhesive strength to dentin. Add pretreatment dissolves the smeariayer, creates a zone of partiaiiy demineralized dentin, and allows the PAA to interact with dentin via the in-termediate layer. Overly aggressive conditioning renders the dentinal tubuies patent, and leaves deeperdemineralized dentin that does not form part of the intermediate layer.

J Adesive Dent 2001.3:153-167 Submitted for publiGatlon:09.10.00: accepted for p\iblicatior\:17.01.01.

^ Honorary Assistant Professor, School of Dentistry, The University ofHong Kong, Hong Kong SAR, China.

" Visiting Research Fellow. Denai School. The University of Adelaide.Austraiia.

' Director of Research. Colgate Australian Ciinicai Dental ResearchCentre, Dental Schooi. The University of Adelaide, Austraiia.

" Professor Emeritus. School of Dentistry, The University of HongKong. Hong Kong SAR. China.

e Reg-enCs Professor. Department of Oral Bioiogy and MaxiiiofaciaiPathology, School of Dentistry. Medical College of Georgia. Au-gusta. Georgia. USA.

Reprint requests: David H. Pashley, Dept. of Oral Biology and Maxilio-faciai Pathoiogy. Schooi of Dentistry. Medical College of Georgia. Au-gusta, Georgia 30912-1129, USA. Tei: **l-706-721-2033. Fax:**l-706-721-6252. e-mail: dpashley@mail.mcg.edu

Conventional chemically-cured (ie, acid-base re-actions) glass-ionomer cements (GICs) v ere

developed almost three decades ago at the Labo-ratory of Government Chemists in England. ^ As arestorative material, GiCs have been successfullyused for the restoration of cervical lesions,^^^^However, poor fracture and wear resistance andlack of radiopacity made early generation chemi-cally-cured GICs problematic as permanent load-bearing restorations.19 This is also true of thesubsequently improved, more viscous formula-tion.^^'^^ High strength GICs that were recentlydeveloped for atraumatic restorative treatment

of posterior teeth exhibit better flexural

Voi 3, No 2, 2001 153

Tay et ai

Strength and produced satisfactory three-year clini-cal results at least for single-surface resto ratio ns>^

To circumvent surface contamination by smearlayers produced during tooth preparation, differentconditioners have been investigated to enhance thelaboratory adhesion of GICs to enamel and dentin,vi/ith equivocai results.2^-29.4i Most manufacturersnow recommend the use of polyacrylic acid (PAA),wih variable concentrations and etching times, asGIC pretreatment protocois. Being a weaii acid. PAAleaves smear plugs intact, i ' ' prevents ingress ofbacteria via exposed dentinai tubuies, and dilu-tion of the setting GICs via transudation of pulpalfluid.^"' Studies that evaiuated the efficacy of PAApretreatment have also produced contradictory re-sults.^'^s xhis may be related, in part, to the thicii-ness of the smear iayers empioyed in differentstudies.2? The controversy is aiso complicated bythe repeated observations of cohesive cement fail-ures instead of interfacial adhesive faiiure in bond-ed GICs.23.37 The low bond strength vaiues of GICsreported using conventional shear and tensiietests''''^'''^ may thus represent the lovj cohesivestrengths of the cements rather than their true ad-hesive strength to dentin.

Aithough GICs bond chemicaliy to tooth substra-tes, their bonding mechanism is complex and is notcompletely understood.^2 on piacement, the hy-drophiiic, aqueous polyalkenoic acid wets the ena-mel and dentin, providing intimate contact with thecement. The proposed mechanism invoives car-boxylic groups of the poiyeiectroiyte replacing phos-phate ions of the substrate to establish ionic bondswith the calcium ions derived from partially dis-soived apatite crystallites.'''' There is aiso initial evi-dence suggesting that chemical bonding of PAAmay also occur with collagen,^^ Ion exchange andreprecipitation of ions released by the glass parti-cles and those dissociated from the part ial iydemineraiized dentin^s.is probably account for theSEM observation of a f luoridated intermediatelayer,^^ an interaction zone,^"^ an interdiffusionzone,^ or the presence of an interphase^^ betweenGiC and dentin. Aithough the glass-hydrogel inter-face in GICs have been documented for some timeusing transmission electron microscopy (TEM),i^ ul-trastructural characterization of the GlC-dentininterface remains elusive. The use of acidic condi-tioners for GIC pretreatment, irrespective of v ihe-ther or not they are rinsed away before bonding,invariably results in some demineralization of theintact dentin. It is not clear hovii the intermediate

layers in water-based GICs differ from microme-chanically retentive hybrid iayers that are producedby contemporary resin-based, seif-etching and totai-etch adhesives.

This study evaluated the uitrastructure of theGlC-dentin interface when a chemically cured (ie,acid-base reaction) GiC was bonded to coronaldentin. To evaluate how GICs interact with the bond-ing substrate, different pretreatment protocols wereused, comprising PAA acids of variabie concentra-tions and etching times and phosphoric acid. Thisproduced a series with progressively aggressiveetching effects on the smear iayer-covered dentin.Uitrastruotures of the GlC-dentin interfaces thatwere tested using the non-trimming version of themicrotensiie bond test^" Vi ere further correlatedwith bond strength. Bonded specimens that werestressed to failure were aiso examined with cryo-SEM to determine whether oohesive failures in GiCwere minimized when the GIC Vi as used for bondtesting. The null hypothesis was that the ultrastruc-ture of the GlC-dentin interface produced by thevarious pretreatment protocols does not correlatewith their tensile bond strengths to dentin.

MATERIALS AND METHODS

Tooth Preparation

Ail bonding was performed on the occlusal surfacesof mid-coronai dentin of 15 extracted, human thirdmoiars from young adults of both genders. Tineywere stored in 0,5% chloramine T at 4°C and usedwithin one month following extraotion. The teethwere prepared by removing all of the occlusal ena-mei using a slow-speed saw with a diamond-impreg-nated disk (isomet, Buehler, Lake Bluff, iL, USA)under water lubrication. A 180-grit silicon carbidepaper v fas used under running water to create asmear iayer on the dentin surface.

Pretreatment Protocols and G/C Bonding

The prepared teeth were randomly divided into fiveexperimental groups v fith three teeth each, basedon the aggressiveness of the pretreatment protocolemployed before the placement of a viscous resto-rative GIC (ChemFlex; Dentspiy De Trey, Konstanz,Germany), The composition of the GIC and the con-ditioners are listed in Tabie 1. In all groups, includ-

154 The Journal of Adhesive Dentistry

Tay et al

Table 1

GIC

ChemFlex

Glass-ionomer cements and conditioners used In this study

Manufacturer

DentsplyDeTreyKonstanz,Germany

ESPE, Seefeld,Germany

Bisco,Schaumburg,IL, USA

Batchnumber

9810000648

FW0046892

0000007171

Conditioner

GIC liquid (10%polyacrylic acid)

Ketac conditioner(25% polyacrylicacid)Uni-Etcn(32% phosphoricacid)

Composition

Powder: Strontiumalumino-fluotosilicate glassPolyacrylic acidTsrtsricacidIron oxide and tita-nium oxide pigments

Liquid: WaterPolyacrylic acid

ing the control specimens, the dentin surfaces werekept wet with water until they were blot dried. Theywere then ieft inverted on a piece of wet, lint-freetissue until the next treatment step.

Group C consisted of three teeth in which the GICwas appiied without any pretreatment of the dentinsurfaces. This served as the control group.

Group P teeth were pretreated with 10% PAA(ChemFlex liquid) for 10 s, and then bonded withoutfurther rinsing.

Group R teeth were pretreated with the Chem-Flex liquid for 20 s. After rinsing the conditionerwith distiiled water for 10 s, each tooth was blot-dried and left inverted on a piece of wet, lint-freetissue untii ready for the placement of the GIC,

Group K teeth were pretreated with 25% PAA(Ketac Conditioner; ESPE, Seefeld, Germany] for 25s. The dentin surfaces were then rinsed for 10 s,then treated as in group R.

Group H teeth were pretreated with 32% phos-phoric acid (Uni-Etch; Bisco. Schaumburg, USA) fer15 s. After rinsing for 10 s, the dentin surfaceswere treated as in group R.

The manufacturer's instructions recommend theuse of a less viscous mix for the first increment andprovide a green scoop that produces a power:iiquidratio of 3.3:1, After applying a 0.5-mm-thick layer ofthis less viscous mixture to all specimens to enablebetter adaptation to the moist, etched dentin, the

rest of the GIC buildup was made using the manu-facturer's recommendation for a more condensableconsistency, using their red scoop that provides apowder:iipuid ratio of 3,8:1, Ail specimens werebuilt up to a height of 5 mm. After the GICs set, aiispecimens were coated with Ketac-Glaze (ESPE) toprotect the set materiai from the loss or gain ofmoisture. Aii teeth were then stored at 37°C, 100%humidity for 24 h before sectioning.

TEM Preparation

One tooth from each group was randomly selectedtor TEM examination. Two 1-mm-thick bondeddentin slabs were vertically sectioned from eachtooth. They were fixed in Karnovsky's fixative (2,5%giutaraldehyde and 2% paraformaldehyde in 0,1mol/L cacodylate buffer, pH 7,3) for 30 mln andrinsed thorough iy with 0,1 moi/L sodium cacodylatebuffer. The specimens were not post-fixed so thatfurther elemental analyses couid be done afterTEM examination. The undemineralized siabs werethen dehydrated in an ascending ethanol series(30% to 100%), immersed in propylene oxide as atransition fluid, and embedded in epoxy resin (TAAB812 resin, TAAB Laboratories, Aidermaston, UK) at60°Cfor48 h.

The epcxy resin-embedded slabs from each

Vol 3, No 2, 2001155

Tay etai

group were sectioned into 1 x 1 x 2 mm blocks thatinciuded the undemineralized dentin on one sideand the GIC on the other. These blocks were thenre-embedded in epoxy resin to facilitate subsequentpreparation of thin sections. Seventy- to 90-nm-thick undeminerali^ed sections were prepared withan uitramicrotome (Reichert iJltracut S, Leica, Vi-enna, Austria) using a diamond knife (Dlatome, Bi-enne, Switzeriand), and coiiected on 200-meshcopper grids (TAAB laboratories, Aldermaston, UK),The unstained sections were examined using atransmission electron microscope (Philips EM208S,Eindhoven, The Netherlands) operating at 80 kV,

Microtensile Bond Testing

The other two specimens in each group were indi-vidually secured with sticky wax to a Plexiglass sec-tioning block. Using the nontrimming techniquedeveloped by Shono et al,3i beams with cross-sec-tional areas of approximately 0,81 mm^ were pre-pared, with the GIC comprising the upper half of thebeam and dentin comprising the lower half. Eachtooth contributed equally, yielding between 20 to23 beams for bond testing. Each beam was fixed toa modified Bencor Multi-T testing assembly (Dan-ville Engineering, San Ramon, CA, USA] using cya-noacrylate (Zapit; DVA, Corona, CA, USA), Thebeams were pulled to failure under tension using auniversal testing machine (Model 4440; Instron,Canton, MA. USA) at a crosshead speed of 1 mmper min.32 The exact dimensions of each fracturedbeam were then individually measured using a digi-tal caliper (Model CD-6BS; Mitutoyo, Tokyo, Japan),from which the tensile bond strength was calcu-lated. The failure modes of the bonds were initiallyevaluated at 30X with a stereo microscope. Failureswere classified as either adhesive between dentinand the GIC, cohesive failure within the GIC, ormixed (combinations of cohesive failure in the GICand adhesive failure along the dentin surface).

One reason for using the microtensile bond testis to reduce the number of teeth required for inves-tigations involving multiple variables, since humanteeth are in short supply. Although only 3 teethwere bonded per group, and one was divided into 2specimens for TEM examination, the other 2 teethper group yielded 20 to 23 beams for bond testingthat were pooled. As the sample size was small, weelected to use a very conservative nonparametricstatistical analysis. Bond strength data obtained for

the five experimental groups were analyzed with anonparametric test (Kruskal-Wallis one-way ANOVA)using a statistical package (SigmaStat Version2.03, SPSS, Chicago, IL, USA). Statistical signifi-cance was set in advance at the 0,05 level, whilemultiple comparisons were then performed usingthe Dunn's post-hoc test.

Fractograptiic Analysis

The dentin sides of four randomly selected frac-tured beams in each group were examined usingiow temperature SEM, to minimize cracking of anyresidual GIC during specimen preparation for con-ventional SEM observation. The beams were kept at100% humidity until the time of examination. Theywere secured with Zapit in holes that were pre-drilled in a copper stub. The entire stub was thenplunged into liquid nitrogen for 2 min. After trans-ferring to the low temperature preparation chamber(Cyroprep LT7400, Flsons, East Sussex, United King-dom), the stub was maintained at -96''C for 30 minto allow sublimation of excess ice crystals from thefractured surfaces. After coating with gold-palla-dium, the temperature of the specimens was fur-ther reduced to -167°C, The stub was then insertedinto the cryo-chamber of a scanning electron micro-scope (Cambridge Stereoscan 440, Cambridge,United Kingdom] and examined at 10 kV,

RESULTS

Uitrastructure of tiie GIC - Dentin Interfaces

In the absence of any surface pretreatment (groupC), no obvious demineralization was observed onthe surface of the intact intertubular dentin (Figl a ] . Smear layer remnants were embedded in adark gray amorphous interphase that was well dif-ferent iated from the surrounding lighter graypolyalkenoate matrix of the GIC (Fig lb ] . The inter-mediate layer of the GlC-dentin interface, in thisspecimen, resided completely within the smearlayer.

In the specimen that was pretreated with 10%PAA for 10 s (group P), the dentin was partially de-mineralized to a depth of 0.5 um to 0.8 Mm (Fig 2a),A 0,3- to 0,5-Mm-thick gray amorphous intermediatelayer was present on top of this partially demineral-ized zone, and was demarcated from the GIC matrix

156 The Journal of Adhesive Dentistry

Tay et al

M

Fig l a A low-magnification TEM micrograph taken from anunstained, undemineraiized section of the GiC-dentin inter-face in the control group C (no pretreatment). Smear layerremnants (S) are present along the surface of the undeminer-alized, intact dentin (U). Giass particies (G) are found withinthe poiyalkenoate matrix (M) of the GiC. Eiectron-dense, circu-lar bodies (biack arrowheads) are also observed. They proba-bly present incompletely reacted components of the GICmixture. Separation of the GIC from the bonded substrate isevident along the interface (arrows). Bar= 500 nm.

Fig I b A higin-magnification unstained TEM micrograph fromanother location of the specimen shown in Fig l a . The smearlayer is present in the fcrm cf a thinner, more compact por-tion (between black arrowheads) that is about 1 |jm thick entop of the undemineralized intact dentin (U), and a thicker su-perficial portion (arrow) that is very loosely dispersed withinthe polyslkenoste matrix |M) of the GIC. Electron-dense, hee-dle-shaped and plate-like remnant apatite crystallites fromttie smear layer are embedded within a gray amorptiousphase. We speculate that this phase represents the morpho-logical manifestation of the interaction between calcium icnsreleased by the partially demineralized apatites from thesmear layer and the carboxyl groups of the pciyalkenoic acidin the setting GIC. Together, they constitute the intermediateiayer (IL¡ cf the GlC-dentin interface, which in this controlgroup resides completely within the smear layer. G: glass par-ticle; U; undemineralized dentin. Bar = 200 nm.

(Fig 2b). These features were more prominent asthe depth of demineraiization ¡noreased (Fig 3a) ¡nthe specimen treated with 10% PAA for 20 s (groupR). This resulted in the formation of a distinct inter-phase aiong the GiC-dentin interface.^3 A highermagnification of the unstained, undemineralizedsection from this specimen (Fig 3b) that had notbeen exposed to osmium (which is commoniy usedduring processing of specimens for electron mi-croscopy) revealed the presence of collagen band-ing along the junction of the intermediate layer andthe partialiy demineralized zone. They could be dis-tinguished from the segregated, eiectron-denseremnant apatite crystallites that were aiso presentwithin the partiaily demineralized coliagen fibrils.The observation of coliagen banding aiong the sur-face of the partialiy demineralized dentin was un-usual, as the section was not positively stained.

Examination of the specimen pretreated with

25% PAA for 25 s (group K) showed that the depthof demineraiization of the dentin increased toabout 1.5 |jm, with the surface coiiagen fibrils com-pieteiy demineralized (Fig 4a). The position of theintermediate layer appeared to have shifted down-ward, and was located predominantiy within thecompietely demineralized region of the intertubuiardentin (Fig 4b). Faint coliagen banding couid alsobe recognized. However, there were also regions ontop of the partially demineralized zone in which thecollagen fibriis became "invisibie." These wereareas where the unstained coiiagen fibrils werecompletely denuded of their apatite components.This diffusion gradient was even more noticeabie inthe specimen treated with 32% phosphoric acid{group H). A zone of completeiy demineralized coiia-gen was created (Fig 5a) that was approximately 8|jm thick, iviore electron-dense regions with distinc-tive collagen banding could be observed around the

Vol 3, No 2, 2001157

Tay etai

Fig 2a A low magnification unstained TEM micrograph of theGiC-dentin interface in group P (10% PAA 10 s, unrinsed). Thesmear layer is almost completely dissoived and the intact in-tertubular dentin (U) is partially dissoived by the polyacryiicacid, resulting in a partially demineralized zone that is O.S jjmthick. A circumferential layer of peritubular dentin (Pd¡ is pre-sent around the dentinal tubule, in which smear plug rem-nants are present (arrowhead). M: poiyalkenoate matrix of theGIC. Bar= l | j m .

Fig 2b A high-magnification unstained TEM micrograph of anarea similar to that shown in Fig 2a. Remnant apatite crystal-lites are found within the zone of partiaiiy demineralizeddentin (P]. On top of this zone is a 200-nm-thick, gray interme-diate iayer (iL] that can be distinguished from the overiyirgGIC matrix (M). We speculate that this phase, similar to that inFig l b , represents the morphological manifestation of the in-teraction of calcium ions from the partially demineralizeddentin with the PAA. It is possible that this phase also ex-tends into the thin layer of partially demineralized dentin thatis created by PAA pretreatment. Eiectron-dense structures (ar-rowheads) within this intermediate layer are probably rem-nants of the smear iayer, as they are not observed in tfiegroups that were conditioned with more aggressive protocois.Strontium aluminofluorosilicate glass particles can be identi-fied within the GIC matrix. A large, typicai giass particieconsists of a centrai electron-dense glass core (C) and a peri-pheral siliceous hydrogel layer (H). U: undemineraii2ed inter-tubular dentin. Bar - 300 nm.

M TL P' i

Fig 3a A low-magnification unstained TEM micrograph of theGlC-dentin interface in group R (10% PAA 20 s, rinsed). A 500nm tbicK gray, amorphous intermediate layer (IL) is weil de-marcated from the GIC matrix (M), and is located well abovethe 1.2'fj'Ti-thick partiaiiy demineralized zone ¡P). U: undem-ineralized intertubular dentin; T: dentinal tubule. Bar = 300nm.

Fig 3b A higher magnification of Fig 3a. Eiectron-dense ap-atite crystailites are readiiy identified within the partiaiiy de-mineraiized zone (P), Along the junction of this zone and tdeintermediate iayer (IL), banded coilagen fibriis (arrowhead)can also be seen. The observation of banding is unusual, asthe section was not positively stained (see aisc Fig 5b¡.'M:GIC matrix; U: undemineralized intertubular dentin Bar = 300nm.

158 The Journal of Adhesive Dentistry

Tay et ai

Fig 4a A iow-magnifioation unstained TEM micrograph of theGlC-dentin interface in group K (25% PAA 25 s, rinsed). Themore aggressive pretreatment protocoi resuits in a 1.5-pm-thiok demineraiized zone, the superficial portion of whioh iscompletely demineraiized (C) and the basai portion partiallydemineraiized (P]. Siihouettes of upright-positioned coliagenfibriis can be identified aiongthe dentin surface (arrowhead).Dentinai tubules are rendered patent and contain glass parti-óles around the tubular orifices (arrow) as well as the poiyai-kenoate matrix (M) within the tubules, Pd: undemineraiizedperitubular dentin; U: undemineraiized intertubuier dentin.Bar = 300 nm.

Fig 5a A low-magnifioatlon unstained TEM micrograph of theGiC-dentin interface in group iH (32% phosphoric aoid 15 s,rinsed). This aggressive conditioning protocoi produced acompieteiy demineraiized zone ¡C] that was 8 ijm thick. Denti-nai tubules (T) were rendered completely patent, but are notolearly demonstrated in tiiis micrograph due to the cutting an-gulation. A gradient of grayness can be identified, being mostnoticeable aiongthe dentin surface and the lateral walis ofthe dentinai tubuie (arrows). These regions are the first to bedemineraiized by the penetrating phosphoric acid oonditioner,and the ooliagen fibriis shouid thus be compieteiy devoid ofhydrosyapatite orystallites. However, distinctive banding canbe discerned within the coiiagen fibrils at higher magnifica-tion (Fig 5b). The gray regions probably represent the interme-diate layer [IL] that is formed between the ions from the GiCfiliers dissolved in the infiitrating PAA reacting with the de-mineraiized dentin matrix. There is probably a iimiting size forpoiyelectroiytes with high molecular weights that restricts howfar they can diffuse through the bed of demineraiized colia-gen. Incomplete infiltration of the PAA probabiy results in theformation of a 1- to 1.5-ijm-thick intermediate iayer within thesuperficiai part of the completely demineraiized dentin. M:GIC matrix; U; undemineraiized intertubular dentin. Bar = 1

Fig 4b A high-magnification TEM micrograph of an area simi-iar to that depicted in Fig 4a. A gray, amorphous intermediateiayer (iL] can be identified mostiy within the compieteiy de-mineraiized portion of the demineraiized zone. Simiiar to Fig3b, faintiy banded coliagen fibriis can be seen within thisiayer (arrowhead). In contrast, remnant apatite crystailiteswithin the underlying partiaiiy demineraiized zone (P) are seg-regated and more electron-dense. As the depth of deminerai-izaticn is deeper than the previous groups, ion exchangebetween the setting GIC and the partiaiiy demineraiized coila-gen fibrils couid result in the formation of an intermediateiayer that resides predominantiy within the superficial iayer ofcompieteiy demineraiized coiiagen instead of aiong thedentin surface. We speculate that the eleotron-iuoent regionin this unstained section indicated by the asterisii probablyrepresents collagen fibrils that are compieteiy denuded of ap-atite orystaliites, but are too deep to be infiltrated by the PAAfrom the setting GiC, so that no interaction products were de-posited within the coiiagen fibrils and/or interfibrillar spaces.iVl; GIC matrix; U: undemineraiized intertubular dentin. Bar =300 nm.

Voi 3, No 2, 2001 159

Tay et al

Fig 5b A higher magnification of another section after de-mineralization and staining with lead citrate and uranyl ac-etate that react with the PAA to form electron-dense deposits.This demonstrates that the PAA does not diffuse into thedemineralized region more than 1 to 2 [jm. Bar •^ 1 |jm.

Fig 5c A higher magnitication of Fig 5a illustrating the un-usual banding patterns (arrowheads) that are present withinsome completely demineralized, unstained collagen fibrils (C)aiong the dentin surface, Interfibrillar spaces are also infil-trated by a material of similar electron density. M: GIC matrix,Bar = 300 nm.

Fig Ga An unstained, undemineralized micrograpn üiov. ngthe presence of "seeds" (arrowheads) witnin the nydrogellayer (H) of a strontium aluminotluorosilicate glass particle.Each electron-lucent "seed" has an electron-dense peripherythat is connected via pinwheel-like, electron-dense extensionsto the center. They vary substantially in size from particle toparticle. However, they are of a similar size range within a sin-gle glass particle. The composition and function of these"seeds" are currently unknown, C: glass core: M: polyalke-noate matrix of the GIC, Bar - 300 nm.

Fig 8b Another unstained section showing the presence ofmuch smaller "seeds" within a glass particle. The roughly par-allel, electron-lucent spaces (arrowhead) are chatters createdin the section that are caused by cutting of the diamond hnifethrough a brittle material, C: glass core; H: siliceous hydrogellayer; M: polyalkenoate matrix of the GIC, Bar = 500 nm.

dentin surfaces and the lateral wafis of the dentinaltubuies. These regions were the first to be deminer-alized by the penetrating pinosphoric acid condi-tioner, and the coiiagen fibriis were completelydevoid of apatite crystallites. Incomplete infiltration

of the PAA resulted in the formation of a 1- to 1,5-nm-thick intermediate layer within the superficialpart of the completely demineralized dentin. Thiscan be seen in a demineralized, stained section(Fig 5b), A higher magnification of the coiiagen

160 The Journal of Adhesive Dentistry

Tay et al

Table 2 Microtensile bond strength (MTBS) of the high-strength glass-lonomer cement bonded to sound dentin using various pretreatmentprotocols

Group Number of beams tested MTBS (MPa} •

20(1)23(0)22(0)21(0)22 (0|

7,2 ± 1,7 a14,0 ± 3,714,0 ± 3,415,0 ± 2,415,313,2

t SQ. Groups lOentified wilh the same su pi:arenltieses iriflicate the number of specir

•Vaines are mi0.05), The vaiutiley couid beti

Group designations C: control - no pretreatment^ P; 10% poiyacrylic acid (PAA| 10PAA 20 s, rinsed: K: 25% PAfl 25 s, nnseû; H: 32% oriosplioric acid, 15 s, rinsed.

ptsara not significsruiy different (p<that faileû during preparation before

banding along the dentin surface from an unstain-ed, undemineralized section is shown in Fig 5c.

Within the polyalkenoate matrix of the GIC, sur-face-reacted glass particles were readily identified.They possessed an electron-dense core and a pe-ripheral electron-lucent, siliceous hydrogel layer (Fig6a), Numerous "seeds" were consistently observedwithin the hydrogel of these glass particles. The"seeds" varied substantially in size from glass parti-cle to glass particle ¡Fig 6b), However, they were ofa similar size range within the same glass particle.

Microtensiie Bond Strengtiis and FractograpiiicAnaiysis

Mean microtensile bond strength (IJTBS] for the fiveexpérimentai groups are shown in Table 2, Thebond testing revealed that the mean IJTBS for eachof the four groups varied from 7.2 to 15.3 MPa.One-way ANOVA based on ranks (Kruskal-Wallis) in-dicated a significant difference among the groupstested. The Dunn's multiple comparison test showedthat the four GIC groups bonded with dentin pre-treatment were not significantly different from oneanother in terms of mean bond strength. They were,however, significantly higher (p < 0,05) than thecontrol group, in which the GIC was bonded tosmear layer-covered dentin.

SEM examination of the fractured interfacesshowed that adhesive failure was the predominantfailure mode in group C (Rg 7a). A partially peeled-off layer along the dentin side of a fractured sur-

face (Fig 7b) corresponded with the intermediatelayer that incorporated the smear remnants in theTEM observation. In contrast, mixed failure was thepredominant failure mode in the other four groups(Fig 7c), At higher magnification, an Intermediatelayer was discernable between the surface of theintact dentin and the fractured GIC in groups P andR (Fig 7d), The intermediate layer that was shownby TEM to be located within the demineralizeddentin could not be seen in groups K and H, sinceSEM observations only reveal surface details.Patent dentinai tubules were identified from thesurface of the fractured dentin (Fig 7e), Figure 8summarizes the results in a schematic form.

DISCUSSION

There are very few TEM studies of GICs or glassionomer-based adhesives available in the litera-ture.i5,is In particular, ultrastructural reports of theinteraction between chemically cured GIC anddentin are lacking due to the difficulty involved inthe preparation of GlC-dentin interface for TEM ex-amination. Unlike resin-dentin interfaces, the poly-alkenoate matrices of GICs are soluble in acidicbuffers or chelating agents^-^ and are rapidly dis-solved during laboratory demineralization of thespecimens. Our pilot TEM investigation of these fiveexperimental groups using stained, demineralizedsections revealed severe dissolution of the matri-ces. Moreover, the glass particles were also dis-solved and became empty shel ls^" that were

Vol 3, No 2, 2001 161

Tay et al

Fig 7a A low-magnification SEM micrograph cf the dentinside of a representative fractured beam from control group C(no pretreatment) that was stressed to failure using the mi-crotensiie bond test. The failure mode is adhesive in nature.At this magnification, a thin, partially peeled-off layer (pointer)can be discerned on the fractured surface, leaving the dentin(D) exposed in some areas. Bar = 20 gm.

Fig 7b A higher magnification of Fig 7a showing that the par-tially peeled-off layer (arrows) corresponds with the TEM ap-pearance of an intermediate layer (IL¡ that resides completelywithin the smear layer. D: intact dentin. Bar = 10 [¡m.

Fig 7c A low-magnification SEM micrograph of the dentinside of a fractured beam from group R (10% PAA 20 s,rinsed). A mixed failure mode can be seen, with a combina-tion of adhesive failure along the surface of the intact dentin(D) and cohesive failure within the GIC (G). This mixed failuremode is also representative to groups P, K and H. Some voids(V) can be discerned within tiie fractured GIC. Cracks are pre-sent on the surface of the exposed GIC even when examinedusing cryo-SEM. They are probably caused by dehydration ofthe thin layer of fractured GIC wheh the fractured beam wasexamined with s stereomicroscope to determine the faiiuremode. IL: intermediate layer. Bar = 20 Jm.

Fig 7d A higher magnification of Fig 7c, showing the pres-ence of an intermediate layer (IL) between the fractured GIC(G) and intact dentin (D|. Bar = 10 pm.

162 The Journal of Adhesive DenUstry

Tay et al

Fig 7e A SEM micrograph showing a higher magnification ofa representative mixed failure mcde in group H (32% phos-phoric acid 15 s, rinsed]. Exposed dentinal tubules (pointers)are readily discerned from the surface of the intact dentin (D).isolated isiands of fractured GIC (G) can also be identified. Anintermediate layer cannot be recognized on top of the intactdentin. Based on the previous TEM observations, the interme-diate layer is probabiy represented by the exposed surface ofthe intact dentin. Bar = 10 ijm.

Fig S Schematic summarizing the depth ofdemineralization cf dentin by the variousconditioners and the depth of the intermedi-ate layer within the underlying dentin.

ControlNo

treatment

10 sec10% PAAunrinsed

20 sec10% PAA

rinsed

25 sec25% PAA

rinsed32% H3PÜ,

rinsed

surrounded by artifactual surface precipitation ofelectron-dense amorphous deposits (Tay. unpub-lished results). There is further evidence that unsi-lanized fluoroaluminosilicate glass particles, whenused as fillers for experimental composites, arecompletely dissoived by the uranyl acetate (pH 4.2)used for staining of undemineraiized TEM sections(Tay, unpubiished results). Therefore, we elected toreport the ultrastructure of the GlC-dentin inter-faces using unstained, undemineralized sections.These sections are much more difficult to preparethan undemineralized dentin bonded using resin-modified GICs or resin-based adhesives. This is be-cause of the lack of resin support and the difficuity

of the hydrophobic epoxy resin to infiltrate the hy-drophilic poiyalkenoate matrix of the GIC.

TEM examination of undemineralized, unstainedspecimens of GIC bonded to dentin has expandedour understanding of the structure of GIC as well asits interaction with the bonding substrate. The pres-ence of "seeds" within the hydrogei layer of thegiass particles in GIC has been previously report-ed^^ but not discussed. They are not unique toChemFiex, as similar "seeds" are also in the fiuo-roaluminosiiicate giass particles of a giass-ionomerbased, ail-in-one dentin adhesive known as React-mer Bond (Shofu, Kyoto, Japan; Tay et ai, unpub-lished results). Wilson's group explained that G-200

Vol 3, No 2, 2001163

Tay et al

glass in the giass-ionomer cement has tv io phases;a continuous calcium aluminosilicate matrix andpartly crystalline calcium fluoride-rich droplets, thenature of which depends on the thermal history ofthe glass.i The setting process of the cement takesplace when the glass is mixed with polyacryiic acid.It has two overlapping reaction stages correspond-ing to a rapid leaching of calcium ions from thenoncrystailine portion of the droplets, follovned by aslovifer release of aluminum (and some calcium)from the major glass phase.^ These processes areaffected by the microstructure and microcomposi-tion of the glass.^

Cur TEM results also confirmed the existence ofan interphase along the GlC-dentin interface, as re-ported by previous SEÍV1 studies. This intermediatelayer was shown in a recent spectroscopic study tocontain elements from the bidirectional diffusion ofions between the GiC and dentin.^^ ip the presentstudy, the intermediate iayer could be iocated ei-ther within the smear layer, on the surface of thepartially demineraiized dentin, or even within thesuperficial part of the completely demineraiizeddentin ¡ie, around the interfibrillar spaces. Fig 8).The inclusion of either smear layer remnants orbanded collagen fibrils within the intermediate layermay be expiained by the aggressiveness of differentpretreatment protocols in removing the smear iayerand demineralizingthe underlying intact dentin. Wespecuiate that the unusual observation of bandedcoiiagen in undemineraiized sections (Fig 5c) maybe caused by the interaction of free metallic ions(0a*2, Sr*2 or Ar^) in the demineraiized coiiagennetwork and the PAA that may adhere chemically toeither remnant hydroxyapatite crystallites'^'' or colla-gen.2i Simiiar to what may occur around the inter-fibrillar spaces, deposition of amorphous, insoluble,organic saits within the gap zones of the collagenmoiecules could result in the formation of vaguebanding within the unstained collagen fibriis.

In the absence of any pretreatment, the interme-diate iayer incorporated only smear remnants, andno demineralization Vi as observed in the underlyingintact dentin (Fig 8). This weak link between thesmear layer and the underlying dentin substratewas ref lected by the signif icant ly lower bondstrength in the control group, and the partial sepa-ration of the intermediate iayer from the underlyingdentin that v as depicted in the fractographic analy-sis, Peutzfeldt and Asmussen^' reported that PAApretreatment improves the bond of GiC to roughdentin surfaces with thici<er smear iayers, but not

for smoother dentin surfaces with thinner smearlayers. In the present study, dentin was abradedwith 180-grit SiC paper, which produces a 0.7- to3,3-|jm-thick smear layer.35 There is the possibilitythat the intermediate layer may not completely in-corporate the smear remnants in areas where thesmear layer is thick. This would be comparabie tothe weakness observed in early generation dentinadhesives that attempted to bond only to smear lay-ers, but without demineralizing and infiltrating theunderiying dentin to form hybrid layers.^

Wesenberg et al reported the presence of a par-tially demineraiized subsurface zone beneath anouter zone of inoreased calcium and phosphoruscontent in unconditioned, GlC-bonded dentin.'^^ Asthe PAA in the GIC mixture exhibits some intrinsicacidity prior to setting,^^ the difference betweenthe observations in our control group and the re-suits of Wesen berg et al may be due to the differentsmear layer thicknesses in the two studies as wellas differences in the concentration of the PAA. Nev-ertheless, in the other four groups that were pre-treated with either PAA or phosphoric acid, weobserved the coexistence of partially or completelydemineraiized subsurface zones beneath the inter-mediate iayers. The variation in the thickness ofthese subsurface demineraiized zones must neces-sarily be related to the aggressiveness of the pre-treatment protocols. There also appeared to be adownward migration (Fig 8) of the intermediate lay-ers from the surface of the partially demineraiizeddentin in the iess aggressive protooois (ie, groups Pand R) to within the superficial portions of the com-pletely demineraiized dentin in the more aggressiveprotocols (ie, groups K and H). Such a reiationshipmay be associated with the limited diffusivity of thePAA and the availability of calcium ions within dif-ferent depths of the demineraiized dentin. It is alsonotable that group P was the only group in whichthere was the iikelihood that the intermediate layercorresponded with the depth of deminerlization.

Intermediate layers of GiC that incorporated in-tact, banded collagen fibrils may be comparable tothe interdiffusion zones (hybrid iayers) in acid-condi-tioned, resin-infiltrated dentin,^^s Although theirmechanisms of formation are different, both thePAA and resins have to diffuse downward through abed of demineraiized collagen fibrils in order to in-filtrate the interfibriliar spaces for either chemicaladhesion or micromechanical retention. This diffu-sion gradient^'' Vi as readily apparent in GIC bondedto phosphoric acid-conditioned dentin. Unlike treat-

164 TheJournai of Adhesive Dentistry

Tay et al

ment with 25% PAA, in which some of the PAA maybe retained by chemical bonding to the partiallydemineralized hydroxyapatites and/or coiiagenafter rinsing, the demineraiized coiiagen in thephosphoric acid group must rely on the infiltrationof PAA from the setting cement. We did not antici-pate the presence of an intermediate layer that con-tains banded collagen in phosphoric acid-etcheddentin, as the apatite crystaliites are completeiydissolved within the demineralized coiiagen net-work. However, there may stiii be residuai calciumions within the interfibrillar spaces of this networkeven after rinsing.^^ j ^ g metáis may also have orig-inated from the action of PAA on the fiiler particles.We are currently investigating the relation betweenintermediate layers and partially/completely de-mineralized dentin in the various GlC-dentin interfa-ces with the use of SElVl/energy dispersive x-ray(STEM/EDX).

Previous studies of dentin adhesives that con-tain polyalkenoic acid copolymer showed that thiscomponent was preferentiaiiy retained on the acid-etched dentin surface, with minimal diffusion (ca500 nm) into the underiying demineralized colla-gen.39 This may be due to the low permeability ofthis high moiecuiar weight resin-modified compo-nent.^" Unlike the diffusion of dentin adhesive com-ponents, such as hydroxyethyl methacrylate, intothe interfibrillar spaces, there is a possibiiity thatthese spaces may not be completely infiltrated byhigh moiecuiar weight PAAs in chemicaiiy cured (ie,acid-base reaction) GICs. The situation may even beworsened by the collapse of the demineralized col-lagen networki'' when acid-etched dentin is desic-cated before GIC placement. Based on our ultra-structural observations, we initially speculated thatthere may be a deterioration of the GlC-dentin bondwith the use of more aggressive pretreatment proto-cols that leave a bed of denuded collagen withinthe subsurface demineralized dentin. Our microten-sile bond strength data, however, did not supportsuch a speculation. As there was no significant dif-ference in the four pretreated groups, we have toaccept the null hypothesis that uitrastructure of theGIC interfaces do not correlate with their respectivemicrotensiie bond strength.

Although the version of microtensile bond testemployed in this study was discriminative enoughto distinguish the unconditioned control group fromthe pretreated groups, it did not allow isolation ofany difference in the latter four groups. When thefracture modes were further analyzed, most failures

were mixed failures and cohesive faiiures in theGIC in these four groups. Thus, the similar bondstrengths in these groups (14,0 to 15,3 MPal repre-sent the cohesive strength of the GIC tested ratherthan its true adhesive strength to dentin. In the twoaggressively pretreated groups (K and H), it is possi-ble that the GIC is bonding to denuded collagen viathe intermediate iayer. If this hypothesis is correct,then the absence of cohesive faiiure within thedemineralized zones suggests that the adhesivestrength of the GIC to dentin is below the ultimatetensile strength of demineralized dentin collagen,which was reported by Sano et al to be in the rangeof 30 MPa,31

It may also be concluded that bonding of Chem-Fiex GIC to dentin may be effectively aciiieved withthe use of mild dentin conditioning protocols suchas the 10% PAA, The use of more aggressive condi-t ioning protocois that demineralize dentin to adepth beyond that necessary for chemicai interac-tion via the formation of the intermediate layerdoes not enhance the dentin-GIC restorative bondstrength. One of the potential disadvantages of amore aggressive acid pretreatment protocol is thatremovai of the smear plugs may increase the per-meability of dentin and outward flux of dentinal fiu-ids during bonding.^ This may further dilute thechemically cured GiCs, and produces weaker bondsto dentin that are simiiar to the effects of dentinperfusion on the bonding of resin-modified GICs,25Another potential disadvantage is the presenceof denuded coiiagen beneath GlC-bonded dentin,which has to be further investigated and confirmed.However, even if this situation exists, it may notbe easily detected ciinicaliy if both the cohesivestrength of the GIC and the bond of the GIC to den-tin are weaker than the uitimate tensile strength ofdemineralized coiiagen. Only when collagen fibrilscan be seen on the GIC side of a failed bond couldone interpret the failure as being due to degrada-tion of the naked collagen fibriis. Although someclinicians advocate the total-etch technique forplacing conventional GICs, we are concerned tiiatthis technique etches dentin deeper than the inter-mediate iayer can form, making the durabiiity ofsuch restorations questionable. Clearly, the inter-phase between dentin and ionomers must be inves-tigated further using other giass-ionomer cementsand related materiais.

Vol 3, No 2, 2001 165

ACKNOWLEDGiVIENTS

We thank Amy Wong and W,S, Lee of the Eleetron MicroseopyUnil, The University of Hong Kong, fur llteir teehnieal ,<upport. Thisstudy was supported hy grant DE Q6427 fioiii the National Instituteof Dental and Craiiiofaeial Researeh, USA. The GIC used in thisstudy was generously supplied hy Dentsply Asia (Hong Kong). Theauthors are grateful lo Michelle Burnside lor secretarial suppori.

REFERENCES

1, Barry Tl, Clinton DJ, Wilson AD, The structure of a glass-iono-mer eement and its reiationsliip to the setting process. JDent Res 1979;58:1072-1079.

2, Berry EA, Powers JM, Bond strength of glass ionomers tocorcnai and radicuiar dentin, Oper Dent 1994:19•122-126.

3, Bertaccliini SM, Abate PF, Blank A, Baglieto MF, Macchi RL.Solubility and fluoride release in ionomers and compomers.Quintessence Int 1999:30:193-197,

4, Csttani-Lorente MA, Godin C, Meyer JM, Early strength ofglass ionomer cements. Dent Mater 1993;9:57-62,

5, Ciucchi B, Bouiilaguet S, Holz J, Pashiey D. DentinsI fluid dy-namics in numan teefh, in vivo. J Fndod 1995:21:191-194.

6, Eiok JD, Gwinnett AJ, Pashiey DH, Robinson SJ, Current con-cepts on adhesion to dentin, Crit Rev Oral Biol Med 1997:8:306-335.

7, Ewoldsen N, Covey D, Lavin M, The physical and adhesiveproperties of dental cements used for atraumatic restorativetreatment. Spec Care Dent 1997:17:19-24,

S. Ferrari M, Davidson CL, Interdiffuson of a traditional glassionomer cement into conditioned dentin. Am J Dent 1997;10:295-297,

9. Fcrss H Release of fluoride and other elements from light-cured glass ionomers in neutral and acidic conditions, J DentRes 1993:72:1257-1262.

10. FranKenberger R, Sindel J, Kramet N, Viscous glass-ionomercements: a new alternative to amalgam in the primary denti-tion'' Quintessence Int 1997;25:667-676,

11. Frencken J, Mahoni F, A treatment technique for tooth decayin deprived communities. World Health 1994:47:15-17,

12. Frencken JE, Makoni F, Sithole WD, ART restorations andglass ionomer sealants in Zimbabwe: Survivai after 3 years.Community Dent Oral Epidemioi 1998:26:372-381,

13. Geiger SB, Weiner S, Fiuoridated carbonatoapatite in the in-termediate iayer between glass ionomer and dentin. DentMater 1993:9:33-36,

14. Gwinnett AJ, Chemically conditioned dentin: a comparison ofconventional and environmental scanning electron microsco-py findings. Dent Mater 1994:10:150-155,

15. Hatton PV. Brook JM, Characterisation of the ultrastructure ofglass-ionomer ¡poly-alkenoate} cement, Br Dent J 1992:173:275-277.

16. Hewlett ER, Caputo AA, Wrobet DC, Glass ionomer bondstrength and treatment of dentin with poiyacrylic acid. J Pros-thet Dent 1991:66:757-772.

17. Ho TR, Smales RJ, Fang TSD, A two year clinical study of twoglass ionomer results used in the atraumatic restorativetreatment (ART) technique. Community Dent Oral Epidemiol1999:27:195-201.

18, Inoue S. Van Meerbeek B, Vargas M, Yoshida Y, LambrechtsP, Vanherle G, Adhesion mechanism of self-etching adhe-sives. In' Tagami J, Toledano M, Prati C (eds). Proceedings ofConference on Advanced Adhesive Dentistry, Third Interna-tional Kuraray Symposium, December 3-4, 1999, Granada,Soain, Cirimido, Italy: Grafiche Erredue, 2000:131-148.

19, Naasan MA, Watson TF, Conventional glass ionomers as pos-terior restorations: A status report for the American Journalof Dentistry, Am J Dent 1988:11:36-45,

20, Ngo H, Mount GJ, Peters MC. A study of glass-ionomer ce-ment and its interface with enamel and dentin using a low-temperature, high-resolution scanning electron microscopictechnique. Quintessence Int 1997:28:63-69,

21, Nezu T, Winnik FM, Interaction cf water-soluble collagen withpoly(acrylic acid), Biomaterials 2000:21:415-419.

22, Nicholson JW. Chemistry of glass-ionomer cements: a review,Biomaterials 1998;19:485-494.

23, 0ilo G, Bond strength of new ionomer cements to dentin,Scand i Dent Res 1981:89:344-347.

24, PasHiey DH, Ciucchi B, Sano H, Horner JA, Permeability ofdentin to adhesive agents. Quintessence Int 1993:24:618-631.

25, Perdigao J, Van Meerbeek B, Lopes MM, Ambrose WW. Theeffect of a re-wetting agent on dentin bonding. Dent Maler1999:15:282-295,

26, Pereira PN, Sano H, Ogata M, Zheng L, Makajima M, TagamiJ, Pashiey DH. Effeot of region and dentin perfusion on bondstrengths of resin-modified glass ionomer cements, J Dent2000:28:347-354.

27, Peutifeldt A, Asmussen E, Effectof poiyacrylic acid treatmentof dentin on adhesion of glass ionomer cerrent. Acta CdontclScand 1990:48:337-341.

28, Powis DR, Folieras T, Merson SA, Wilson AD, Improved adhe-sion of a glass ionomer cement to dentine and enamel, JDent Res 1982:61:1416-1422,

29, Prati C, Nucci C, Montanari G, Effects of acid and cleansingagents on shear bond strength and marginal microleakage ofglass-ionomer cements. Dent Mater 1989:5:260-265,

30, Radin S, Ducheyne P, Falai;e S, Hammond A, In vitro transfor-mation of bioactive glass granules into Ca-P shells, J BiomedMater Res 1999:49:264-272,

31, Sano H, Takatsu T, Ciucchi B, Russell CM, Pashiey DH, Ten-sile properties of resin-infiltrated deminerali;ed humandentm, J Dent Res 1995:74:1093-1102,

32, Schreiner RF, Chappell RP, Glaros AG, Eick JD, Microtensiletesting of dentin adhesives. Dent Mater 1998:14:194-201,

33, Sennou HE, Lebugle AA, Grégoire GL X-ray photoeiectronspeotroscopy study of the dentin-giass ionomer cement inter-face. Dent Mater 1999:15:229-237,

34, Shono Y, Ogawa T, Terashita M, Carvalho RM, Pashiey EL,Pashiey DH, Regional measurement of resin-dentin bondingas an array, J Dent Res 1999:78:699-705,

35, Tay FR, Sano H, Carvalho R, Pashiey EL, Pashiey DH, An ultra-structural study of the influence of acidity of self-etchirgprimers and smear layer thickness on bonding to intactdentin, J Adhes Dent 2000:2:83-98.

36, Tyas MJ, The effect of dentine conditioning with polyacylicacid on the clinical performance of glass ionomer cement - 3year results, Aust Dent J 1994;39:220-227,

166 The Journal of Adinesive Dentistry

Tay et al

37. Tyler MW, Fitzgerald M, Dennison JB, Heys DR. The effect ofpulpal fluid flew on tensile bond strength of a glass-ionomercernent; an in vivo and in vitro comparison. Oper Dent1994:19:116-120-

38. Vah Dijken JW. Fcur-year evaluation of the effect of 10% poly-acrylic acid cr water rinsing pretreatment on retention ofglass Dolya I ken oate cement. Eur J Oral Sei 1996:104:64-66.

39. Van Meerbeek B, Conn U, Duke ES, Eick JD, Robinson SJ,Guerrero D. Correlative transmission electron microscopy ex-amination of n onde minerai i zed and demineraUzed resin-dentin interfaces formed by two dentin adhesive systems. JDent Res 1996:75:879-888.

40. Walls AW, McCabe JF, Murray JJ. Factors influencing the bondstrength between giass poiyalkenoate (ionomer) cementsand dentine. J Oral Rehabil 19S8;15:537-547.

41. Weiger R, Heuchert T, Hafin R, Lost C. Adhesion of a glassionomer cement to human radicular dentine. Endod DentTraumatol 1995:11:214-219.

42. Wesenberg G, Hals E. The in vitro effect of a glass ionomercement on dentine and enamel walls. An electron probe andmicroradiographic study. J Oral Rehabil 1980;7:35^2.

43. Wilson AD, Kent BE. A new translucent cement for dentistry.The glass ionomer cement. Br Dent J 197 2; 132:133-135.

44. Yoshida Ï, Van Meerbeek B, Nakayarra V, Snauwaert J, Helle-mans L, Lambrechts F, Vanherie G, Wakasa K. Evidence ofchemical bonding at biomaterial-hard tissue interfaces. JDent Res 2000:79:709-714.

Vol 3, No 2, 2001 167

Lab Manual of Normal

Oral Histology

HollisLon L.Riviere

128 pp (spiral binding}215 color illustrationsISBN 0-86715-386-5US$34/£22

This lab manual of color oral histologyphotographs is designed to aid dental anddental hygiene students in identifying thesalient features of microscopic anatomy oforal ti.ssue.s. It presents large, cicar, identi-fiable photos of normal oral tissues anddeveloping teeth accompanied by briefdescriptions highlighting the special fea-tures of each.

Contents

Chapter 1 Tooth DevelopmentChapter 2 EnamelChapter 3 DentinChapter 4 PulpChapter 5 CementitmChapter 6 Perioduntal Ligament and

Den togingival JunctionChapter 7 Alveolar BoneChapter 8 Tooth Ertiption and SheddingChapter 9 MucosaChapter 10 Salivary GlandsChapter 11 Temporomandibular Joint

To Order

CaU ToU Free 1-800-621-0387or Fax 1-630-682-3288

Quintessence Pitbhshing Co, Inc

Visit our web sitehttp://www.qumtpub.com