How Can Nanoleakage Occur in Self-etching Adhesive Systems

Vol 4, No 4, 2002 255

How Can Nanoleakage Occur in Self-etching Adhesive Systems that Demineralize and

Infiltrate Simultaneously?

Franklin R. Taya/Nigel M. Kingb/Kar-mun Chanc/David H. Pashleyd

a Honorary Assistant Professor, Pediatric Dentistry and Orthodon-tics, Faculty of Dentistry, The University of Hong Kong, Hong KongSAR, China.

b Professor, Pediatric Dentistry and Orthodontics, Faculty of Dentist-ry, The University of Hong Kong, Hong Kong SAR, China.

c Graduate Student, Pediatric Dentistry and Orthodontics, Faculty ofDentistry, The University of Hong Kong, Hong Kong SAR, China.

d Regent’s Professor, Department of Oral Biology and MaxillofacialPathology, School of Dentistry, Medical College of Georgia, Augus-ta, Georgia, USA

Purpose: Single-step adhesives which etch and prime simultaneously and are not rinsed should not ex-hibit areas of incomplete infiltration within hybrid layers produced in sound dentin. This study examinedthe extent of silver uptake using ammoniacal silver nitrate in three two-step, self-etching primers (Imper-va Fluoro Bond, Shofu; UniFil Bond, GC, ABF system, Kuraray) and one single-step, self-etching adhesive(AQ Bond, Sun Medical) bonded to dentin and four poly(HEMA) resins used as controls.

Materials and Methods: Flat dentin surfaces were bonded with these adhesives and sectioned into0.8-mm-thick slabs that were then coated with nail varnish except for the bonded interfaces and im-mersed in AgNO3 for 24 h. Four types of poly(HEMA) resins were made: 100% HEMA; 90% HEMA-10%water; 75% HEMA-10% water, all polymerized with TBBO at 50°C for 6 h; 100% HEMA polymerized at25°C for 30 min. After developing, undemineralized, unstained, epoxy resin-embedded sections wereprepared for TEM.

Results: Nanoleakage patterns were observed in all bonded specimens. Fine segregated silver particlesand reticular silver-staining patterns were found within the thin hybrid layers created by the threeself-etching primers. For the single-step, self-etching adhesive, heavy silver deposits were identified with-in the hybridized complex formed by this adhesive within the smear layer, the underlying intact dentin,and in the adhesive layer. Increasing amounts of silver uptake were observed in poly(HEMA) specimenscontaining more water or that were polymerized at 25°C for a short time instead of 50°C for 6 h.

Conclusions: Silver uptake in hybrid layers formed by self-etching adhesives in sound dentin is not nec-essarily caused by disparities between the depths of demineralization and resin infiltration. They repre-sent areas of increased permeability within a polymerized resin matrix in which water is incompletelyremoved resulting in regions of incomplete polymerization and/or hydrogel formation.

J Adhes Dent 2002; 4: 255–269. Submitted for publication: 14.05.02; accepted for publication: 07.08.02.

Reprint requests: Dr. Franklin Tay, Pediatric Dentistry and Orthodon-tics, Faculty of Dentistry, The University of Hong Kong, Prince PhilipDental Hospital, 34 Hospital Road, Hong Kong SAR, China. Tel:+852-2859-0251, Fax: +852-2559-3803. e-mail: [email protected]

ontemporary dentin adhesives are classifiedinto three-step, two-step and single-step sys-

tems, depending on how the three cardinal stepsof etching, priming, and bonding to tooth sub-strates are accomplished or simplified.14 Two-stepsystems may further be subdivided into theself-priming adhesives that require a separateetching step, and the self-etching primers that re-quire an additional bonding step.25 Self-etchingprimers and single-step, all-in-one adhesives areattractive in that removal of the smear layer andsmear plugs is not required. This reduces the po-tential for postoperative sensitivity,2,23,41 as wellas the bonding problems associated with transu-

C

Tay et al

256 The Journal of Adhesive Dentistry

dation of dentinal fluid through patent dentinal tu-bules.15 The technique sensitivity associated withbonding to a dehydrated collagen matrix is alsoeliminated,31,37 as water is an essential compo-nent in these systems for effective demineraliza-tion of dental hard tissues.38

Self-etching adhesives vary in aggressiveness byvirtue of the differences in composition and con-centration of polymerizable acids and/or acidic res-in monomers in these systems. The milder ver-sions can etch beyond clinically relevant smear lay-ers to form thin hybrid layers in intact dentin.40 Themore aggressive systems completely dissolve thesmear layer and create hybrid layers that are al-most as thick as those found in phosphoric ac-id-etched dentin.38 In the absence of adjunctiveacid etching,12,24 hybrid layers that are formed us-ing self-etching adhesives, regardless of their thick-ness, should not exhibit areas of incomplete infil-tration as the bonding substrates are simulta-neously demineralized and infiltrated by the sameresin components.43

Nanoleakage was originally used to describe mi-croporous zones beneath or within hybrid layersthat permitted tracer penetration. Unlike microleak-age that occurred through gaps between compos-ites and hybrid layers, nanoleakage occurred in theabsence of interfacial gaps. It occurs through sub-micron-sized spaces within dentin hybrid layerswhere disparities existed between the depths ofdemineralization and monomer diffusion.33,34 Al-though nanoleakage studies of resin-dentin inter-faces bonded using self-etching primers werescanty, they showed silver uptake into thin hybridlayers formed by some self-etching systems.20,36

As the samples in these two studies were pro-cessed 24 h after bonding, it is unlikely that thetracer-penetrated microporous zones were causedby long-term degradation of resin-dentin bonds.9,35

There is a possibility, however, that minerals suchas amorphous calcium phosphates that are re-pre-cipitated in the bonded interfaces of these nonrins-ing adhesives may be dissolved after immersion inan acidic silver nitrate (AgNO3) solution (pH =3.4).21 This may produce artefactual microporosi-ties that give rise to false positive results that leadto spurious conclusions.6

This transmission electron microscopical (TEM)study examined the extent of tracer penetrationinto the bonded interfaces of dentin and four con-temporary self-etching systems using a modifiedsilver-staining technique. To avoid the possibility of

inadvertent mineral dissolution with an acidicAgNO3 solution, a basic version of AgNO3 (pH =9.5) was used in this study.19,39 Addition of ammo-nium ions into the AgNO3 solution produces chelat-ed ionic complexes: the diamine silver ions (alsoknown as silver diamine; [Ag(NH3)2]+). The null hy-pothesis tested was that silver uptake within thinhybrid layers produced by self-etching adhesives insound dentin are not always caused by disparitiesbetween the depths of demineralization and resininfiltration.

MATERIALS AND METHODS

Tooth Preparation and Bonding Procedures

Eight recently extracted human third molars wereused for the experiment. They were stored in 0.5%chloramine T at 4oC and used within one month ofextraction. The teeth were prepared by first remov-ing the occlusal enamel using a slow-speed sawwith a diamond-impregnated disk (Isomet, Buehler,Lake Bluff, IL, USA). A 180-grit silicon carbide (SiC)paper was used under running water to create asmear layer of clinically relevant thickness on thedentin surface.40 Bonding was performed on theocclusal surfaces of mid-coronal dentin. Four non-rinsing, self-etching adhesives were examined, withtwo teeth used for each adhesive. Three of themare two-step, self-etching primers that require addi-tional placement of a layer of bonding resin to cou-ple the resin composite to the primed dentin. Thefourth one is a single-step, single bottle self-etch-ing adhesive with an additional adhesion promoterincorporated in the application sponge. The majorcomponents and batch numbers of these adhe-sives are shown in Table 1.

Group FB (Imperva Fluoro Bond, Shofu, Kyoto,Japan)Imperva Fluoro Bond is a two-step, fluoride-releas-ing self-etching adhesive system. The mixedself-etching FB-primer A and B components wereplaced on the dentin with agitation for 30 s. Withoutrinsing, the primer was gently air dried with oil- anddust-free air (Dust Off Plus, Falcon Safety Products,Branchburg, NJ, USA) to remove the primer solvent.FB-bond was then applied and light cured for 10 swith a halogen light-curing unit (Variable Intensity Po-lymerizer, Bisco, Schaumburg, IL, USA), with the cur-ing intensity set at 500 mWcm-2. A lining resin com-

Tay et al

Vol 4, No 4, 2002 257

posite containing colloidal silica particles andpre-polymerized fillers (Protect Liner F, Kuraray, Os-aka, Japan) was applied to facilitate subsequentTEM specimen preparation. The composite was add-ed in three 1-mm increments, each light cured for40 s.

Group UF (UniFil Bond; GC, Tokyo, Japan) UniFil Bond is a two-step self-etching adhesive sys-tem. The single-bottle self-etching primer was ap-plied to the polished dentin surface with agitationfor 20 s. Without rinsing, the primed surface wasgently air dried to remove the primer solvent. A lay-er of filled bonding agent was then applied to theprimed dentin, gently thinned, and then light cured

for 10 s. The lining composite was applied in a sim-ilar manner.

Group AB (ABF experimental system; Kuraray,Osaka, Japan)ABF is an experimental antibacterial, fluoride-re-leasing, self-etching adhesive. The single-bottleself-etching ABP primer was applied on the pol-ished dentin with agitation for 20 s. Without rins-ing, the primed surface was gently air dried to evap-orate the primer solvent. A layer of KBF bondingresin was then applied to the primed dentin, gentlyair thinned and then irradiated for 10 s. This wasfollowed by a similar application of the lining com-posite.

Table 1 Self-etching primer/adhesive systems examined in the study

System Composition Batch number

Imperva Fluoro Bond (Shofu, Kyoto, Japan)

FB-Primer A:Water,a,b catalyst

090062 (FB-primer A)090078 (FB-primer B)090072 (FB-Bond)(two-step fluoride-releasing

self-etching primer)FB-Primer B:Acetone,a HEMA,b 4-AET, 4-AETA, catalyst

FB-Bond17% silanated PRG-Ca fillers, 3% silanated colloidal silica, 4-AET, HEMA,b UDMA,b TEGDMA, photoinitiator

UniFil Bond(GC, Tokyo, Japan)

Self-etching primer:Water,a,b ethanol,a 4-MET, HEMA,b UDMA,b photoinitiator

9906171

(two-step self-etching primer) Bonding agent:HEMA,b UDMA,b TEGDMA, silanated colloidal silica

ABF(Kuraray, Osaka, Japan)

Antibacterial primer (ABP):Water,a,b MDPB, MDP, HEMA,b dimethacrylates, photoinitiator

000411 (ABP)991130 (KBF)

(experimental two-step, antibac-terial, fluoride-releasing self-etching primer)

Bonding agent (KBF)MDP, HEMA,b dimethacrylates,bsilanated colloidal silica, surface-treated sodium fluoride crystals, photoinitiator

AQ Bond(Sun Medical, Shiga, Japan)

AQ Bond base liquid:Water,a,b acetone,a 4-META, UDMA,b HEMA,b MMA, photoinitiator

VK5

also marketed by Parkell (Farm-ingdale, NY, USA) as Touch&Bond

AQ sponge:Polyurethane foam, p-toluenesulfinic acid sodium salt (p-TSNa)

(single-step, self-etching adhesive)

a: adhesive solvents. b: ingredients common to all adhesives. Abbreviations: 4-AET: 4-acryloxyethyltrimellitic acid; 4-AETA: 4-acryloxyethyltrimellitate anhydride; 4-MET: 4-methacryloxyethyltrimellitic acid; 4-META: 4-methacryloxyethyltrimellitic anhydride; HEMA: 2-hydroxyethyl methacrylate; MDP: 10-methacryloyloxydecyl dihydrogen phosphate; MDPB: 12-methacryloyloxydodecylpyridinium bromide; MMA: methyl methacrylate; PRG-Ca: calcium type pre-reacted glass ionomer fill-ers; UDMA: urethane dimethacrylate; TEGDMA: triethylene glycol dimethacrylate.

Tay et al


Group AQ (AQ Bond, Sun Medical, Shiga, Japan)This single-step, single-bottle, light-curable self-etch-ing adhesive is also marketed by Parkell (Farm-ingdale, NY, USA) in North America as Touch&Bond.

One piece of AQ sponge was placed in a dis-posable dispensing dish included in the kit. Twodrops of AQ Base liquid were then added to satu-rate the sponge with the Base liquid that was lib-erally applied to the polished dentin surface for20 s. This first coat was gently air dried for 3 to5 s and then light activated for 5 s. A second coatof the activated Base liquid was applied with thesame AQ sponge, dried evenly for 5 to 10 s, andthen irradiated again for 10 s, using the same vis-ible light-curing unit. The lining composite was ap-plied in the same manner as the other three adhe-sives.

The bonded teeth of all groups were stored indistilled water at 37°C for 24 h.

Preparation of Ammoniacal Silver Nitrate

Twenty-five grams of silver nitrate (AgNO3) crystals(Sigma Chemical, St. Louis, MO, USA) were dis-solved in 25 ml of 28% aqueous ammonium hydrox-ide (NH4OH; Sigma Chemical) in the presence ofambient laboratory light. This created a suspensionof black silver particles. Additional 28% NH4OH wasused to titrate the black solution, with magneticstirring at room temperature (25 ± 2°C), until thesolution slowly became clear as the ammoniumions complexed the silver into diamine silver ions([Ag(NH3)2]+). This solution was diluted to 50 mlwith distilled water to achieve a 50 wt% solution(pH = 9.5) that was transferred to an amber bottlefor storage.

Silver Staining Technique

After storing in distilled water at 37°C for 24 h,two 0.8-mm-thick slabs were prepared from eachtooth by sectioning perpendicularly to the bondedinterface with the use of the Isomet saw under wa-ter lubrication. One of the two slabs harvestedfrom each tooth was coated with two layers offast-setting nail varnish (Lancaster, Berlin, Germa-ny) applied 1 mm from the bonded interfaces.Without allowing these slabs to be dehydrated,they were immersed immediately in the ammonia-cal silver nitrate solution in the dark for 24 h. We

did not allow the tooth slabs or varnish to dry com-pletely so as to avoid the creation of artefactualsubmicron hiati beneath the resin-dentin interfac-es that could be mistaken for nanoleakage (Tayand Pashley, unpublished results). The other slabserved as a control and was not exposed toAgNO3, but remained immersed in distilled waterfor the same 24 h.

The silver-stained slabs were rinsed six times indistilled water and placed in photodeveloping solu-tion for 8 h under a fluorescent light to facilitate re-duction of the silver ions into metallic silver parti-cles within potential voids along the bondedinterfaces46 as well as in the polymerized resins.22

TEM Preparation

The silver-stained slabs were fixed in Karnovsky’sfixative (2.5% glutaraldehyde and 2% paraformalde-hyde in 0.1 mol/L cacodylate buffer, pH 7.3) for 1 hand rinsed thoroughly with 0.1 mol/L sodium ca-codylate buffer. Two 1-mm-wide rectangular beamswere sectioned from each slab across the bondedinterface. Without further postfixation or demineral-ization, the beams were dehydrated in an ascendingethanol series (30% to 100%), immersed in propy-lene oxide as a transition fluid, and embedded in ep-oxy resin (TAAB 812 resin, TAAB Laboratories, Alder-maston, UK) at 60oC for 48 h. Each resin-embed-ded beam was trimmed down into a 2 x 1 mm blockthat contained the silver-stained interface andre-embedded in epoxy resin. Sixty 100-nm-thick, un-demineralized sections were prepared using an ul-tramicrotome (Ultracut S, Leica, Vienna, Austria)and a diamond knife (Diatome, Bienne, Switzer-land). The sections were collected on single slot,carbon- and formvar-coated copper grids (ElectronMicroscopy Sciences, Fort Washington, PA). With-out further staining, the silver-stained interfaceswere examined with a transmission electron micro-scope (Philips EM208S, Eindhoven, The Nether-lands) operating at 100 kV.

To provide an overall view of the resin-dentin in-terface of each adhesive without exposure to silvernitrate, the second slab from each tooth was com-pletely demineralized in an aqueous solution of 0.1mol/L ethylene diamine tetra-acetic acid (EDTA)that was buffered with sodium hydroxide to a pH of7. The slabs were fixed, dehydrated and resin em-bedded as described previously. Seventy-nm-thicksections were collected using 200-mesh copper

Tay et al

Vol 4, No 4, 2002 259

grids (TAAB laboratories, Aldermaston, UK). Theywere stained with 2% uranyl acetate for 10 min, fol-lowed by Reynold’s lead citrate for 7 min and exam-ined with the same microscope at 80 kV.

Preparation and Silver Staining of Poly(HEMA)

As a control, poly(HEMA) specimens of similar di-mensions were created to determine how much sil-ver would penetrate into the resin in the absence ofa resin-dentin interface. Pure HEMA was used as itis a common ingredient of bonding formulationswhose exact composition remains a proprietary se-cret. Hydroxyethyl methacrylate (HEMA) can poly-merize in the presence of water to form flexible wa-ter-insoluble solids known as hydrogels.30 Four dif-ferent versions of poly(HEMA) were prepared for sil-ver staining and TEM examination. In group 1, 10ml of 100% HEMA (Sigma Chemical) was chemical-ly polymerized in a Teflon mould (1.5 x 1.5 x0.5 cm) by the addition of 10 drops of tri-N-butylbo-rane (TBBO, Parkell, Farmingdale, NY, USA). Withoutthe additional use of a cross-linking resin, the resinmonomer was polymerized at 50°C for 6 h. In group2, a poly(HEMA) hydrogel was prepared by addingthe same amount of catalyst to a solution of 90vol% HEMA and 10 vol% distilled water and poly-merized under the same conditions. In group 3, apoly(HEMA) hydrogel was prepared in a similar wayusing 75 vol% HEMA and 25 vol% distilled water.Group 4 was prepared using 100% HEMA and thesame amount of catalyst, but unlike the other spec-imens that were polymerized at 50°C, this solutionwas left to polymerize at room temperature for30 min only to produce a rubbery, incompletely po-lymerized polymer.

A 1-mm-thick slab was sectioned from the centerportion of the glassy1 well-polymerized 100%poly(HEMA) block (group 1), using the Isomet sawwith the slowest available speed and under waterlubrication. As the poly(HEMA) hydrogels (groups 2and 3) and the incompletely polymerized poly(HE-MA) block (group 4) were rubbery in consistency,1

they were sliced into slabs of similar thickness us-ing a razor blade. These slabs were immersed inammoniacal AgNO3 solution for 24 h, and thenrinsed and developed as previously described. Af-ter silver staining, 2 x 2 mm blocks of poly(HEMA)were produced from each of the four specimens.They were allowed to dry completely and were sim-ply supported with epoxy resin without passing

through ethanol or propylene oxide to prevent fur-ther swelling of the hydrogels and/or dissolution ofincompletely polymerized oligomers.

For specimens of groups 1 and 4, 70 nm thicksections were prepared with a diamond knife usingthe conventional method of sectioning and collect-ed using 200-mesh copper grids. For the two hydro-gel specimens (groups 2 and 3) that remained in arubbery state after processing, the resin blockswere chilled with solid carbon dioxide (dry ice) priorto sectioning to lower the temperature of the hydro-gels to below their glass transition tempera-tures.32 Semithin sections were cut with a dry dia-mond knife. Resin blocks were then re-chilled and70 nm ultrathin sections were quickly preparedwith a wet diamond knife using 10 times the nor-mal speed of cutting to prevent softening of the hy-drogels via water sorption. The sections were re-trieved using single slot, carbon- and formvar-coat-ed copper grids. All TEM sections were examinedat 80 kV without further staining. The size and per-centage distribution of silver particles in each dig-itized micrograph were determined using imageanalysis software (NIH Image 1.60, Scion, Freder-ick, MD, USA).

RESULTS

Dentin Bonded Specimens

Demineralized specimens that were stained withuranyl acetate and lead citrate are shown in Fig 1.Imperva Fluoro Bond (Group FB), UniFil Bond(Group UF), and the ABF experimental system(Group AB) all produced hybrid layers of similarthickness (ca 0.5 to 1 µm) within the intact intertu-bular dentin. The bonded interface in the unfilledAQ Bond (Group AQ) contained numerous blisters,many of which were directly above the tubular orific-es. High magnification views of these systems(Fig 2) showed that the smear layer was completelydissolved in groups FB, UF, and AB. Collagen fibrilson the surface of the hybrid layer were upright andexhibited a shag-carpet appearance. By contrast,the hybridized complex in group AQ was thicker insome areas due to the incorporation of smear layerremnants (ca 1 to 1.5 µm thick) on top of the hy-bridized intertubular dentin (ca 0.5 to 1 µm).

None of the silver-stained undemineralized sec-tions exhibited frank gaps with heavy silver depos-its that would indicate the presence of microleak-

Tay et al


Fig 1a Uranyl acetate and lead citrate-stained, low magnifi-cation TEM micrograph of demineralized resin-dentin inter-faces treated in Imperva Fluoro Bond (group FB). The hybridlayer within the intact intertubular dentin varied between 0.5to 1 µm. The smear layer was completely dissolved. A: adhe-sive layer; H: hybrid layer; D: laboratory demineralized inter-tubular dentin; SP: smear plug; Arrowhead: stained fillers.

Fig 1b Uranyl acetate and lead citrate-stained, low magnifi-cation TEM micrograph of demineralized resin-dentin inter-faces in UniFil Bond (group UF). The hybrid layer within theintact intertubular dentin varied between 0.5 to 1 µm. Thesmear layer was completely dissolved. A: adhesive layer; H:hybrid layer; D: laboratory demineralized intertubular dentin;SP: smear plug.

Fig 1c Uranyl acetate and lead citrate-stained, low magnifi-cation TEM micrograph of demineralized resin-dentin inter-faces treated with ABF, an experimental antibacterialadhesive system (group AB). The hybrid layer within the in-tact intertubular dentin varied between 0.5 to 1 µm. Thesmear layer was completely dissolved. A: adhesive layer; H:hybrid layer; D: laboratory demineralized intertubular dentin;SP: smear plug; Arrowheads: stained fillers.

Fig 1d Uranyl acetate and lead citrate-stained, low magnifi-cation TEM micrograph of demineralized resin-dentin inter-faces in AQ Bond (group AQ). The hybrid layer within theintact intertubular dentin varied between 0.5 to 1 µm. Nu-merous blisters (B) were observed along the resin-dentin in-terface. Smear layer remnants were incorporated as part ofthe hybrid layer in some areas, making it thicker than 1 µm.A: adhesive layer; H: hybrid layer; D: laboratory demineral-ized intertubular dentin; SP: smear plug.

Tay et al

Vol 4, No 4, 2002 261

Fig 2a Group FB, high-magnification view of the stained,demineralized resin-dentin interfaces shown in Fig 1a. Col-lagen fibrils from the surface of the hybrid layers exhibitedupright, shag-carpet appearances (pointer) after completedissolution of the smear layer. Arrowhead: stained fillers. A:adhesive layer; H: hybrid layer within intact dentin; D: labora-tory demineralized intertubular dentin.

Fig 2b Group UF, high-magnification view of the stained,demineralized resin-dentin interfaces shown in Fig 1b. Col-lagen fibrils from the surface of the hybrid layers exhibitedupright, shag-carpet appearances (pointer) after completedissolution of the smear layer. A: adhesive layer; H: hybridlayer within intact dentin; D: laboratory demineralized inter-tubular dentin.

Fig 2c Group AB, high-magnification view of the stained,demineralized resin-dentin interfaces shown in Fig 1c. Col-lagen fibrils from the surface of the hybrid layers exhibitedupright, shag-carpet appearances (pointer) after completedissolution of the smear layer. Arrowhead: stained fillers. A:adhesive layer; H: hybrid layer within intact dentin; D: labora-tory demineralized intertubular dentin.

Fig 2d Group AQ, high-magnification view of the stained,demineralized resin-dentin interfaces shown in Fig 1d. A 1.1to 5-µm-thick remnant smear layer (Sm) was incorporated aspart of the hybridized complex. A: adhesive layer; H: hybridlayer within intact dentin; D: laboratory demineralized inter-tubular dentin.

Tay et al


age along the bonded interfaces. In Group FB, trac-er infiltration was confined to discontinuous is-lands of metallic silver aggregations within the hy-brid layer only (Fig 3a). These silver deposits wereintensely electron-dense (black) compared with theunstained, mildly electron-dense (grey) GIC fillersand colloidal silica clusters that were found withinthe 20-µm-thick adhesive layer. No silver depositswere present within the dentinal tubules or the ad-hesive layer (Fig 3b).

In Group UF, silver staining was confined to the0.5-µm-thick hybrid layer and was absent fromthe 20-µm-thick adhesive layer (Fig 4a). Withinthe partially demineralized hybrid layer, silver de-posits were present as a more or less continuousline of fine, individual silver granules along theentire adhesive-hybrid layer interface (Fig 4b). Insome localized regions, discontinuous, reticular

patterns of heavier silver aggregates were alsoidentified within the entire thickness of the hybridlayer (Fig 4c).

A similar pattern of silver staining was present inthe hybrid layer of Group AB (Fig 5a). The distinc-tion between the interfacial line of fine silver parti-cles and the underlying discontinuous reticular pat-terns of heavier silver aggregates can be clearlydiscerned in Fig 5b. Contrary to the previous twoadhesive systems, silver deposits were alsopresent in the adhesive layer of group AB (Fig 5c).

The extent of silver deposits in the previousthree groups were mild when compared to the se-vere silver staining that was observed in both thehybridized complex and the adhesive layer inGroup AQ (Fig 5c). At a higher magnification, heavysilver deposits were observed along the entirethickness of the hybridized complex, and in the

Fig 3a TEM micrograph of undemineralized, silver-stainedsections from group FB. A low magnification view depictingthe overall extent of silver penetration, that was manifestedas discontinuous islands of electron-dense (black), reducedmetallic silver granules (arrowhead) within the hybrid layeronly. Pointers: mildly electron-dense, silanated PRG-Ca fill-ers that are formed by the complete reaction of fluoroalumi-nosilicate glass particles and polyalkenoic acid. A: adhesivelayer. Between arrows: extent of the hybrid layer. C: lining res-in composite. M: mineralized intertubular dentin.

Fig 3b TEM micrograph of undemineralized, silver-stainedsections from group FB. A high magnification view of the100- to 300-nm-diameter silver deposits within the com-pletely demineralized and partially demineralized zone (P) ofthe hybrid layer. No nanoleakage was detected in the dentin-al tubule or adhesive layer. Colloidal silica clusters (arrow-head) were present within the adhesive layer. Pointer: mildlyelectron-dense, silanated PRG-Ca fillers that are formed bythe complete reaction of fluoroaluminosilicate glass parti-cles and polyalkenoic acid. A: adhesive layer. Between ar-rows: extent of the hybrid layer. M: mineralized intertubulardentin. SP: smear plug.

Tay et al

Vol 4, No 4, 2002 263

Fig 4b A high magnification view of Fig 4a showing thatnanoleakage was manifested as a more or less continuousline of individual metallic silver granules (40-80 nm in dia-meter) along the surface of, and as discontinuous islands ofelectron-dense, silver aggregates within the partially dem-ineralized hybrid layer (between arrows). By contrast, colloi-dal silica clusters (arrowhead) from the adhesive (A) did notappear as electron-dense as the silver deposits. A fine retic-ular pattern of silver deposits could also be found in the un-derlying mineralized dentin (M). This feature was onlyobservable when a section was under 80 nm thick. SP: smear plug. Pointer: mineralized debris that was trappedwithin the the polimerized adhesive.

Fig 4a Undemineralized, silver-stained sections from groupUF. A low magnification view depicting the overall extent oftracer penetration that was manifested as a fine continuousstreak of black metallic silver granules within the hybrid layeronly. C: lining resin composite. A: adhesive layer. Between ar-rows: extent of the hybrid layer. M: mineralized intertubulardentin.

Fig 4c A higher magnification of Fig 4b showing a reticularpattern of silver deposits within the hybrid layer (between ar-rows). Arrowhead: colloidal silica clusters. A: adhesive. M: mineral-ized intertubular dentin.

Tay et al


smear plugs that were present in the dentinal tu-bules (Fig 6a).

Poly(HEMA) Resin Specimens

Figure 7 depicts the distribution of silver particlesin different poly(HEMA) specimens. Sections takenfrom the glassy polymer block produced from 100%HEMA (group 1) revealed fine silver particles scat-tered throughout the entire section (Fig 7a). Thesesilver particles occupied 0.79% of the area of thepolymer matrix. Larger silver particles were identi-fied in the two poly(HEMA) hydrogels (groups 2 and3). The percentage area distribution in the 90 vol%HEMA (group 2) hydrogel (Fig 7b) and 75 vol% HEMA(group 3) hydrogel (Fig 7c) sections were 1.25% and4.24%, respectively. In contrast, a section takenfrom the incompletely polymerized, 100% HEMApolymer block (group 4) revealed heavy silver depos-its that accounted for 39.45% of the matrix space(Fig 7d).

Fig 5a Undemineralized, silver-stained sections from groupAB. A low magnification view depicting the overall extent oftracer penetration within the hybrid layer and the adhesivelayer of this self-etching adhesive. Both mildly elec-tron-dense colloidal silica clusters (arrowhead) and NaF crys-tals (pointer) were present in the adhesive layer. Betweenarrows: extent of hybrid layer. A: adhesive layer. M: mineral-ized intertubular dentin.

Fig 5b Undemineralized, silver-stained sections from groupAB. A high magnification view showing the manifestation ofnanoleakage within the partially demineralized hybrid layer.Very fine (ca 15 nm), individual metallic silver granules couldbe identified along the hybrid layer-adhesive interface. A dis-continuous, reticular pattern of silver aggregates (60 to90 nm diameter) was also observed along the base of thehybrid layer. A: adhesive layer. Between arrows: extent of thehybrid layer. M: mineralized intertubular dentin. P: mineral-ized peritubular dentin. T: dentinal tubule.

Fig 5c A very high magnification view of the NaF crystals(pointer) within the adhesive layer in Fig 5a. Silver deposits(arrowhead) within these crystals probably represent spacesleft behind, following the elution of fluoride ions.

Tay et al

Vol 4, No 4, 2002 265

DISCUSSION

Contemporary self-etching adhesive formulationsusually include acidic hydrophilic monomers, water,HEMA, and bifunctional dimethacrylates.8,45 In-crease in the concentration of acidic monomers inthese systems is required for etching through thesmear layer into the underlying dentin11,44. Waterprovides the medium for ionization of these acidicresin components. HEMA is included as a transi-tional polymerizable solvent as many of the acidicmonomers are not readily water-soluble.13 Other bi-functional or multifunctional monomers are includ-ed to provide strength via cross-linking of the poly-mer matrix.47

In this study, we avoided the possibility of resid-ual mineral dissolution within partially demineral-ized hybrid layers by mildly acidic AgNO3 by using a

basic ammoniacal silver nitrate solution. Presum-ably, silver tracer should be absent when no dis-crepancy exists between the depth of demineral-ization and the extent of resin infiltration. However,silver deposits were still observable in all the hy-brid layers examined. In particular, a substantialamount of silver deposits was identified from boththe hybrid layer and the adhesive layer of the sin-gle-step, self-etching adhesive, AQ Bond. Althoughthe relatively high pH (ie, 9.5) of the basic AgNO3may have degraded unprotected collagen, the sil-ver uptake probably represented regions of subop-timal conversion within the polymer matrix due toincomplete removal of solvent.20 It is known thatwater can inhibit polymerization of dentin adhesiveresins,16 and that incompletely polymerized resinshave affinities for specific stains.5 As evident fromFig 7d, the incompletely polymerized resin matrix

Fig 6a Undemineralized, silver-stained sections from groupAQ. A low magnification view depicting the overall, complexpattern of tracer penetration within the entire resin-dentin in-terface. Electron-dense (black) metallic silver deposits werefound within the full thickness of the hybridized complex. Dif-fusion of the tracer into the adhesive layer appeared as an-nulus-shaped deposits around the partially-exposed blistersoverlying the dentinal tubules (arrowhead), as interconnect-ed, thin black lines within the first (A1) and second layer (A2)of the adhesive, and as thick finger-like projections betweenthese two adhesive layers (pointer). C: lining resin compos-ite. Between arrows: extent of the hybridized complex. M:mineralized intertubular dentin.

Fig 6b Underminiralized, silver-stained sections from groupAQ. A high magnification view showing the presence of se-vere nanoleakage (100- to 200-nm-diameter silver particlesclustering into larger particles) within the hybridized complexand the dentinal tubule. A: adhesive layer. Between arrows:extent of the hybridized complex. M: mineralized intertubulardentin; P: mineralized peritubular dentin; SP: smear plug.

Tay et al


Fig 7a TEM micrograph of silver-stained section taken froman initially glassy poly(HEMA) block composet of 100% HE-MA. Microporosities within the polymer matrix (HE) were in-dicated by the presence of electron-dense silver deposits.Tiny silver particles (pointer) were scattered throughout thepolymer matrix. At a higher magnification (insert), these“larger” particles were around 20 nm in diameter. “Smaller”silver granules between 2.5 and 4 nm (arrowheads) in dia-meter were also present.

Fig 7b TEM micrograph of silver-stained section taken froma poly(HEMA) hydrogel consisting of 90 vol% HEMA and 10vol% water. Microporosities within the polymer matrix (HE)were indicated by the presence of electron-dense silver de-posits. Silver particles up to 80 nm in diameter were identi-fied.

Fig 7d TEM micrograph of silver-stained section taken froman incompletely polymerized poly(HEMA) specimen com-posed of 100% HEMA that was polymerized at 25°C for 20min. Microporosities within the polymer matrix (HE) were in-dicated by the presence of electron-dense silver deposits. Aheavy deposition of silver particles up to 150 nm in diameterwas identified. Aggregation of the silver particles suggestedthat the potential microporosities within the incompletely po-lymerized resin matrix were probably interconnected. EX:supporting epoxy resin.

Fig 7c TEM micrograph of silver-stained section taken froma poly(HEMA) hydrogel consisting of 75 vol% HEMA and 25vol% water. Microporosities within the polymer matrix (HE)were indicated by the presence of electron-dense silver de-posits. Silver particles up to 300 nm were identified.

Tay et al

Vol 4, No 4, 2002 267

of pure HEMA permitted extensive silver impregna-tion. In single-step self-etching adhesives, hydro-phobic and hydrophilic resin components are inter-mixed prior to polymerization. Phase separation ofresin components can occur as the solvent inwhich they are mutually soluble – acetone – isevaporated. This probably resulted in blister forma-tion as well as the complex pattern of silver stain-ing within the adhesive layer of AQ Bond. Fraction-ation of a polymer blend in the presence of watermay also explain why silver staining was only ob-served in the adhesive layer of this single-stepself-etching adhesive and not in polymer matrix ofthe other three two-step, self-etching primers.

It is unlikely that incomplete polymerization is re-sponsible for the uptake of silver within the thin hy-brid layers formed by the three self-etching primers.Because of the small molecular size of HEMA,these hybrid layers may largely be composed ofpoly(HEMA) that copolymerize with acidic resinmonomers in the presence of water to form anionichydrogels with pH-dependent swelling behav-iors.17,18 In this study, less complex hydrogel mod-els consisting of poly(HEMA) and water7 were inves-tigated for their permeability to silver ions in the ab-sence of interactions with dentin.

It is known that poly(HEMA) hydrogels are essen-tially porous materials. The dimension of mi-croporosities within the resin matrix depends on thepolymerization techniques.4 Bulk polymerization of100% HEMA results in a transparent, glassy poly-mer that slowly swells in water. Although it allowsthe transfer of some small molecules, this kind ofpoly(HEMA) is considered “nonporous.” The spacesbetween macromolecular chains are the only avail-able loci for the mass transfer, with pore sizes thatare within the range of molecular dimensions,namely several nm (Fig 7a, arrowheads). Pore sizeincreases as HEMA is polymerized in an aqueoussolution. By increasing the amount of water in themonomeric mixture, “microporous” (10 to 100 nm)(Fig 7b) and “macroporous” (100 nm to 1 µm)(Fig 7c) hydrogels can be produced. These hydro-gels are still transparent like the “nonporous” onesand are known as homogeneous hydrogels. Howev-er, when the concentration of water exceeds 45wt%, heterogeneous hydrogels known as poly(HE-MA) sponges are produced, with pore sizes rangingfrom 5 to 80 µm3.

Under normal bonding to smear layer-covered den-tin, the bulk of the adhesive solvents is removed bygentle air drying after the application of self-etching

primers. However, residual water may persist due tolowering of the vapor pressure of water by HEMA,29

resulting in the formation of different forms of ho-mogenous hydrogels. This may produce a hydrogelon the surface of some hybrid layers that permits athin, more or less continuous line of fine silver de-posits as well as the reticular silver-staining patternthat is formed around the collagen fibrils (“mi-croporous” hydrogel). The occurrence of these silverdeposits of variable dimensions attested that hybridlayers that contained polymerized HEMA are perme-able to silver nitrate. Thus, with self-etching primers,the presence of silver deposits within hybrid layersdo not necessarily indicate disparities between thedepths of demineralization and monomer infiltration.It is conceivable that the “nanoleakage“ found in hy-brid layers created by these self-etching systems onsound dentin may actually be a demonstration of thepermeability28 of homogenous hydrogels to waterand ion movement. Hybrid layers composed primarilyof poly(HEMA) are more elastic than those made withbifunctional, cross-linked polymer chains,27 makingthem more effective as stress-absorbing layers dur-ing polymerization shrinkage of the subsequentlybonded resin composites.42 Although porouspoly(HEMA)-containing, cross-linked hydrogels are ofsubstantial clinical value as pharmaceutical sus-tained-release devices for peptides and drugs,30

their rapid drop in physical properties on swelling af-ter water sorption render them less desirable in den-tin bonding.27 This may account for the drop in ten-sile bond strength and loss of resin from hybrid layersin some poly(HEMA)-containing adhesives afterlong-term water storage.10,35

Within the limits of this study, it may be conclud-ed that at least for sound dentin, silver depositspresent within thin hybrid layers produced byself-etching adhesives are not necessarily causedby disparities between the depths of demineraliza-tion and resin infiltration. They are not nanoleakagein the strictest sense, and more likely represent ar-eas of increased permeability within a polymerizedresin matrix in which water is incompletely removedfrom the primed dentin or adhesive layers, resultingin regions of incomplete polymerization and/or hy-drogel formation. These regions may permit higherdiffusional water fluxes within the hybrid layers thatcould accelerate water sorption and the extractionof unpolymerized or degraded monomers. However,it is anticipated that true nanoleakage can occurwhen self-etching adhesives are used in conjunc-tion with acid etching,43 or when they are applied

Tay et al


to caries-affected dentin.26 Tensile bond strengthsof some self-etching adhesives were significantlylower when bonded to caries-affected dentin, al-though the hybrid layers were twice as thick asthose found in sound dentin,26 indicating thatself-etching adhesives were unable to perfectly sealthe considerably more porous, demineralized car-ies-affected and caries-infected dentin. Further re-search should be performed to examine the occur-rence of nanoleakage when self-etching adhesivesare used on these pathological bonding substrates.

ACKNOWLEDGMENTS

We thank Amy Wong of the Electron Microscopy Unit, The Uni-versity of Hong Kong for technical assistance. This study wassupported in part by the Faculty of Dentistry, University of HongKong, and by grant DE 06427 from the National Institute ofDental and Craniofacial Research, USA. The authors are grate-ful to Michelle Barnes for secretarial support.

REFERENCES

1. Andreopoulos AG, Plytaria M. Release of drugs from polymer-ic hydrogels. J Biomater Appl 1998;12:291-299.

2. Brunton PA, Cowan AJ, Wilson MA, Wilson NH. A three-yearevaluation of restorations placed with a smear-layer-mediat-ed dentin bonding agent in non-carious cervical lesions. J Ad-hes Dent 1999;1:333-341.

3. Chirila TV, Chen YC, Griffin BJ, Constable IJ. Hydrophilicsponges based on 2 hydroxyethyl methacrylate, I: effect ofmonomer mixture composition on the pore size. Polymer In-ternational 1993;32:221-232.

4. Crawford GJ, Chirila TV, Vijayasekaran S, Dalton PD, Consta-ble IJ. Preliminary evaluation of a hydrogel core-and-skirtkeratoprosthesis in the rabbit cornea. J Refractive Surg1996;12:525-529.

5. de Gee AJ, ten Harkel-Hagenaar E, Davidson CL. Color dyefor identification of incompletely cured composite resins. JProsthet Dent 1984;52:626-631.

6. Gale MS, Darvell BW. Dentine permeability and tracer tests.J Dent 1999;27:1-11.

7. Gehrke SH, Biren D, Hopkins JJ. Evidence for Fickian watertransport in initially glassy poly(2-hydroxyethyl methacrylate).J Biomater Sci, Polymer Edition 1994;6:375-390.

8. Hasegawa T, Manabe A, Itoh K, Wakumoto S. Investigationof self-etching dentin primers. Dent Mater 1989;5:408-410.

9. Hashimoto M, Ohno H, Kaga M, Endo K, Sano H, Oguchi H.In vivo degradation of resin-dentin bonds in humans over 1to 3 years. J Dent Res 2000;79:1385-1391.

10. Hashimoto M, Ohno H, Kaga M, Endo K, Sano H, Oguchi H.Fractographical analysis of resin-dentin bonds. Am J Dent2001;14:355-360.

11. Hayakawa T, Kikutake K, Nemoto K. Influence of self-etchingprimer treatment on the adhesion of resin composite to pol-ished dentin and enamel. Dent Mater 1998;14:99-105.

12. Igarashi K, Toida T, Nakabayashi N. Effect of Phenyl P/HEMAprimer on bonding to demineralized dentin by phosphoric ac-id. Japan J Dent Mater 1997;16:55-60.

13. Ikemura K, Kouro Y, Endo T. Effect of 4-acryloxyethyltrimelliticacid in a self-etching primer on bonding to ground dentin.Dent Mater J 1996;15:132-143.

14. 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). Proceedingsof Conference on: Advanced Adhesive Dentistry. Third In-ternational Kuraray Symposium. December 3-4, 1999,Granada, Spain. Cirimido, Italy: Grafiche Erredue,2000:131-148.

15. Itthagarun A, Tay FR. Self-contamination of deep dentin bydentin fluid. Am J Dent 2000;13:195-200.

16. Jacobsen T, Söderholm KJ. Some effects of water on dentinbonding. Dent Mater 1995;11:132-136.

17.Khare AR, Peppas NA. Swelling/deswelling of anionic copol-ymer gels. Biomaterials 1995;16:559-567.

18.Kou JH, Amidon GL, Lee PI. pH-dependent swelling and solutediffusion characteristics of poly(hydroxyethyl methacry-late-co-methacrylic acid) hydrogels. Pharmacy Res 1988;5:592-597.

19.Kusanagi A, Yanagibashi K. The ammoniacal silver reactionfor basic nuclear proteins in the spermatids of the mouse.Protoplasma 1975;83:327-338.

20. Li H, Burrow MF, Tyas MJ. Nanoleakage patterns of four dentinbonding systems. Dent Mater 2000;16:48-56.

21. Li H, Burrow MF, Tyas MJ. The effect of load cycling on thenanoleakage of dentin bonding systems. Dent Mater 2002;18:111-119.

22.Mair LH. An investigation into the permeability of compositematerials using silver nitrate. Dent Mater 1989;5:109-114.

23.Mandras RS, Thurmond JW, Latta MA, Matranga LF, KildeeJM, Barkmeier WW. Three-year clinical evaluation of the Clear-fil Liner Bond system. Oper Dent 1997;22:266-270.

24.Miyasaka K, Nakabayashi N. Combination of EDTA condition-er and Phenyl-P/HEMA self-etching primer for bonding to den-tin. Dent Mater 1999;15:153-157.

25.Miyazaki M, Sato M, Onose H, Moore BK. Influence of thermalcycling on dentin bond strength of two-step bonding systems.Am J Dent 1998;11:118-122.

26.Nakajima M, Ogata M, Okuda M, Tagami J, Sano H, PashleyDH. Bonding to caries-affected dentin using self-etching prim-ers. Am J Dent 1999;12:309-314

27.Paul SJ, Leach M, Rueggeberg FA, Pashley DH. Effect of watercontent on the physical properties of model dentine primerand bonding resins. J Dent 1999;27:209-214.

28.Pauly S. Permeability and diffuse data. Chap. VI. In: BrandupJ, Immergut (eds). Polymer Handbook, 3rd edition. Chiches-ter: John Wiley & Sons,1989:435-436.

29.Pashley EL, Zhang Y, Lockwood PE, Rueggeberg FA, PashleyDH. Effects of HEMA on water evaporation from water-HEMAmixtures. Dent Mater 1998;14:6-10.

30.Peppas NA, Bures P, Leobandung W, Ichikawa H. Hydrogelsin pharmaceutical formulations. Eur J Pharm Biopharm2000;50:27-46.

31.Perdigão J, van Meerbeek B, Lopes MM, Ambrose WW. Theeffect of a re-wetting agent on dentin bonding. Dent Mater1999;15:282-295.

Tay et al

Vol 4, No 4, 2002 269

32.Reid N. Cryo-ultramicrotomy - ultrathin sectioning at low tem-peratures. Part II. Section 12.2. Polymers and rubbers. In:Glaubert AM (ed). Practical methods in electron microscopy.Volume 3. Amsterdam: North-Holland Publishing Company,1974:332-338.

33.Sano H, Shono T, Takatsu T, Hosoda H. Microporous dentinzone beneath resin-impregnated layer. Oper Dent 1994;19:59-64.

34.Sano H, Takatsu T, Ciucchi B, Russell CM, Pashley DH.Nanoleakage: leakage within the hybrid layer. Oper Dent1995; 20:18-25.

35.Sano H, Yoshikawa T, Pereira PN, Kanemura N, Morigami M,Tagami J, Pashley DH. Long-term durability of dentin bondsmade with a self-etching primer, in vivo. J Dent Res 1999;78:906-911.

36.Sano H, Yoshiyama M, Ebisu S, Burrow MF, Takatsu T, CiucchiB, Carvalho R, Pashley DH. Comparative SEM and TEM ob-servations of nanoleakage within the hybrid layer. Oper Dent1995;20:160-167.

37. Tay FR, Carvalho R, Yiu CKY, King NM, Zhang Y, Agee K, Bouil-laguet S, Pashley DH. Mechanical disruption of dentin col-lagen fibrils during resin-dentin bond testing. J Adhes Dent2000;2:175-192.

38. Tay FR, Pashley DH. Aggressiveness of contemporaryself-etching adhesives. Part I: depth of penetration beyonddentin smear layers. Dent Mater 2001;17:296-308.

39. Tay FR, Pashley DH, Yoshiyama M. Two modes of nanoleakageexpression in single-step adhesives. J Dent Res 2002;81:472-476.

40. Tay FR, Sano H, Carvalho R, Pashley EL, Pashley DH. An ul-trastructural study of the influence of acidity of self-etchingprimers and smear layers thickness on bonding to intact den-tin. J Adhes Dent 2000;2:83-98.

41.Unemori M, Matsuya Y, Akashi A, Goto Y, Akamine A. Com-posite resin restoration and postoperative sensitivity: clinicalfollow-up in an undergraduate program. J Dent 2001;29:7-13.

42.Uno S, Finger WJ. Function of the hybrid zone as a stress-ab-sorbing layer in resin-dentin bonding. Quintessence Int1995;26:733-738.

43.Walker MP, Wang Y, Swafford J, Evans A, Spencer P. Influenceof additional acid etch treatment on resin cement dentin in-filtration. J Prosthodon 2000;9:77-81.

44.Watanabe I, Nakabayashi N, Pashley DH. Bonding to grounddentin by a phenyl-P self-etching primer. J Dent Res 1994;73:1212-1220.

45.Watanabe I, Nikaido T, Nakabayashi N. Effect of adhesion pro-moting monomers on adhesion to ground dentin. J Japan SocDent Mater Devices 1990;9:888-893.

46.Wu W, Cobb E, Dermann K, Rupp NW. Detecting margin leak-age of dental composite restorations. J Biomed Mater Res1983;17:37-43.

47.Yamauchi J, Masuhara E, Nakabayashi N, Shibatani K, WadaT. Method of filling a tooth cavity. United States Patent num-ber 4,259,075.March 31, 1981.

Documents

How Can Nanoleakage Occur in Self-etching Adhesive Systems