Dent Mater 12:262-271, July, 1996

Morphological field emission-SEM study of the effect of six phosphoric acid etching agents on human dentin

Jorge PerdigHo ly2, Paul Lambrecht#, Bart Van Meerbeek2, &gel0 R. Tom@, Guido Vanherle2, August0 B. Lopes4

‘Institute Superior de Ciencias da Saude, School of Dentistry, Monte da Caparica, PORTUGAL “Catholic University of Leuven, Department of Operative Dentistry and Dental Materials, BELGIUM

3University of Coimbra, Department of Biochemistry and Center for Neurosciences, PORTUGAL 4Uniuersity ofrlveiro, Department of Ceramics and Glass, PORTUGAL

ABSTRACT Objectives. This study evaluated the effects of six phosphoric acid- etching agents on dentin, the independent variables being two acid concentrations (10% and 32%-37%) and three thickener conditions (no thickener, silica, and polymer). The tested hypothesis was that the use of different etchants with similar concentrations of phosphoric acid would result in similar depths of dentin demineralization. Methods. Thirty dentin disks were obtained from extracted human teeth by microtome sectioning. The dentin surfaces were etched with one of the etching agents, fixed, dehydrated and dried. The specimens were observed using a FE-SEM. The mean deepest demineralization of intertubular dentin was measured from the fracture surfaces of the disks. These values were analyzed by ANOVA and Duncan’s test. The morphological appearance of the dentin surfaces was compared using the following observation criteria: 1) Presence of a cuff of peritubular dentin; 2) Relative thickness of the layer containing residual collagen or smear layer particles; and 3) Formation of a submicron hiatus at the bottom of the exposed collagen network. The pH of each of the etching agents was measured. A correlation analysis was made of the pH vs. the depth of dentin demineralization. Results. Silica-thickened etchants did not demineralize dentin as deeply as did polymer-thickened etchants and unthickened etchants. High magnifications revealed three distinct zones within the demineralized dentin layer: an upper porous zone of residual smear layer or denatured collagen and residual silica particles (in groups etched with silica- thickened etchants), an intermediate area with randomly oriented collagen fibers, and a lower zone with a submicron hiatus, few collagen fibers, and scattered hydroxyapatite inclusions. This hiatus was observable in all the specimens etched with the polymer-thickened etchants, in 90% of the specimens etched with the unthickened phosphoric acid liquids, and in 60% of the specimens etched with the silica-thickened gels. Significance. The results obtained suggest that similar concentrations of phosphoric acid etchants containing distinct thickeners result in different demineralization depths as well as different morphology of etched dentin.

INTRODUCTION Scanning electron microscopy (SEMI has been widely used to study the mechanisms of enamel bonding (Gwinnett and

262 Perdig& et aL/SEM of acid-efched denbn

Matsui, 1967) and more recently, dentin bonding (Nakabayashi et al., 1982; Van Meerbeek et al., 1992). However, a detailed study of structures involved in dentin bonding, such as the resin-dentin interditfusion zone, requires high magnification and good resolution, which used to be possible only with transmission electron microscopy (TEM) (Eick, 1992;Van Meerbeeket al., 1993a). Field Emission SEM achieves higher resolution than conventional thermo-ionic SEM due to its brighter electron source and narrower electron beam, so one can work at lower accelerating voltages without compromising resolution (Goldstein et al., 1981; Delong, 1993).

Phosphoric acid is used in dentistry as an etching agent for enamel (Buonocore, 1955) and more recently, for dentin (Fusayamaet al., 1979). It is believed that the depth of dentin demineralization is directly related to the concentration of the acid (Chiba et al., 1989; Pashley et al., 1992).

The interaction of the etching agents with dentin is limited by the buffering effect of hydroxyapatite and other dentin components (Wang and Hume, 1988). The acidic agents remove the smear layer and the super&&l part of the dentin, open the dentin tubules, demineralize the dentin surface, and increase the microporosity of the intertubular dentin (Van Meerbeek et al., 1992; Pashley et al., 1993; Sano et al., 1994). The penetration of the acids occurs primarily along the tubules (Selvig, 1968). Bonding to dentin is thought to basically rely on a micromechanical entanglement of hydrophilic resins into this demineralized microporous dentin, thus forming a reticular intertwined hybrid tissue composed of collagen, residual mineral particles, and resin (Nakabayashi et al., 1982;Van Meerbeeket al., 1993a; 199313).

In spite of having been marketed as liquids for years, most of the current etching agents are now gels, of either thick or thin consistency (Guba et al, 1994). Manufacturers add thickeners to their gels in order to facilitate handling. The advantages of the gel forms are that the clinician can easily control the spread of the acid over the surface and visually identify the presence of the acid (Ruse and Smith, 1991; Guba et al., 1994). ‘I’he demineralizing effect is clinically observed when gas bubbles accumulate within the gel. Etching gels thickened with silica microparticles leave a particulate residue

0.2 M sodium cacody- late buffer at pH 7.4 for 1 h with three changes, followed bv

Aqueous 10% phosphoric acid, 0.48 Private laboratory distilleh water for- obtained from a standard DhosDhoric acid solution 1 min. Theywere then

3 All-Etch/l 19303 10% phosphoric acid with polymer** 0.48 Bisco Dental Products, Itasca, IL, USA

Ultra-Etch 35%/0208%1 FKO 35% phosphork acid with silica Ultradent Products Inc.

dehydrated ir; ascend- ing grades of ethanols (25% for 20 min, 50% for 20 min, 75% for 20 min, 95% for 30 min, and 100% for 60 min). After the final ethanol step, the specimens were dried bv immersion in hexa- methyldisilazane KCH3)3SiNHS~~CH,L, HMDS, Electron Mi- croscope Sciences)] for 10 min, placed on a filter paper inside a covered glass vial, and

on the dentin surface that cannot be completely removed even with vigorous rinsing> but polymer-thickened gels leave a clean dentin surface &er rinsing (Kubo et al., 1991; Perdiggo et al., 1994a). As a result of acid-etching dentin, a layer of denatured collagen and residual smear layer particles. may form on the de&in surface and prevent the collagen network f?-om being codpletely exposkd (Pashley et aE., 1993). It has not been es$blished whether the .silica component of some etching agents partially prevents the formation of this residual superficial layer. Moreover, the effects of other components of e&&g gels, such as surfactants, polymers, and other proprietary modifiers, tie not &lly understood.,

“Tlie hypothesis to be tested in this study was that similar acidic concentrations of tierent phosphoric acid etching agents resdt in similar demineralization depths as we& as a similar morphology of the etihed dentin, regardless of the presence and type of thickener in the etchant.

Thirty extracted caries-free and unrestored human molars storedin an aqueotis solution ofO.5%.chloramine at 4-C within 1 mon of extraction were used in this study The occlusal enamel was removed, ana’30 dentin disks (600 -i: 100 +nn thick) were obtained ,from mid~dle dentin by iamond-saw microtome sect+fiing (Leit? rl600, WetzIar, Germany) parallel to the occlusal surface.‘ The bottom surface of each $sk was coated witi.two layers of&lv+sh. A t&nsversal groove was made on tile varnished,$de of e&h $entin &sk tb facilitate splitting after specimen p&paration (Boj7de &d Wood, 1969). A smear layers was cre&d’ on the top s&face by wet-sanding with 600-$rit ‘Sic ,sanc$ap& for 60 s (Pashley et al, 1988). The spec@eins tyer& ?alr;ldoniny divi+d,tit~ six equal’groups (n = 5) (see Table 1 for sO&ce of ‘a+&$. T+e, phos&oric acid e$hant was kpplied for 15 s, rinsed 0% v@h water for 10 s, and the etched dentin stiace was,,driqd, tith oil-f?ee compressed air foi- 1-k s: The s&wens yere immediately immersed in 2.5% glut&&@hyd~ :Fn’ ‘0.1 M~~~+&m cacodylate &&er (Electron M&$&cope S,~~~ces,.~o~~as~~o~, PA, USA] at pH 7.4 for 12 h & 4°C. After &ation, ihc A;+ &ere rinsed with 20 mL of

air-dried at room temperature (Pe&iggo et al., 1995). After drying, the disks were fractured by carefully applying

bending force on the groove of the;non-etched surface (Boyde andWood, 1969). The disks were mounted on aluminum stubs (Agar Scientific Ltd., Stansted, UK) with carbon cement (Conductive Carbon Cement after GBcke, Neubauer Chemikalien, Mtister, Germany1 and colloidal s&&r paint (Quick Drying Silver Paint,Agar Scientific, Ltd.). One half of each disk was mounted to observe dentinal tubules in cross section and the other halfto observe dent&J tubules longitu- dinally The specimens were coated with gold-palladium by means of a Polaron E-5000 sputter-coater (Polaron Equipment Ltd., Watford, En&and) at 10 mA for 1 min and observed under a Hitachi S-4100 Field Emission Scanning Electron Microscope (Hitachi, Tokyo, Japan) at an accelerating voltage of 6-8 kv a working distance of 16 mm, from magnifications of 2, 500x to 40,000x.

Morphological Eualuation Criteria. The mean deepest demineralization of intertubular dentin (pm) from &actured specimens was measured on the microscope monitor and on photographic prints. The entire interface between demineralized dentin and unaffected dentin was screened in each specimen under a lateral view. The morphological appearance ofthe dentin surfaces was comparatively assessed using the following observation criteria: 1) The ‘cuff’ of superficial peritubuldr dentin observable from an occlusal perspective. This cuff is a circulx rim of peritubular dentin that remains after three-dimensional circumferential demineralization from b&h khe intratubular and ‘the intertul+r sides, which is not observed in deeply etched specimens (Perdiggo et al., 1994b); 2) Tkie relative thickness of the layer containing residual collagen or smeti layer particles (Pashley et al, 1993), or siiica remnants (Perdigao et al.,

1994a), which were observed on the surface corn a lateral perspective. Three different categories were attributed:“thick”, when the residual material obviously occluded the intertubular porosity; “thin”, when the intertubular porosity was mostly patent; and “absent”, when no residual smeared surface layer could be detected, and 3) !I’he formation of a

Dental Maferials/July 1996 263

Fig. 1. Dentin etched with aqueous 10% phosphoric acid (Group 1). a) Cross-sectioned tubules showing intense intertubular microporosity, exposed outer peritubular collagen fibers, and peripheral intertubular collagen fibers. Note fibrous structures inside the tubules (arrows) which may correspond to remnants of odontoblastic processes. Bar ??3 urn. b) Tubule entrance containing a fibrous structure. The intertubular dentin is less porous around the tubule (arrow). Bar ??750 pm. c) Lateral view of the exposed collagen network and the transition between demineralized and unaffected dentin with a small hiatus (white arrows). Collagen fibers are embedded in mineral (asterisks); residual smear layer particles are present (black arrow), and pores are observed. D - unaffected dentin. S - dentin surface. Bar = 750 urn. d) Another lateral view of Fig. lc showing the transition between peritubular and intertubular dentin (oval). Residual mineral particles were lefl within the exposed collagen network (asterisk). Bar = 750 urn.

submicromillimeter hiatus at the bottom of the exposed collagen network (Perdigao et al., 1995). This submicro- millimeter space has been described as being located at the transition between the densely packed exposed collagen fibers and the unaffected dentin.

revealedsignificant differences (at p < 0.0001) in mean intertubular demineralization depth based on two independent variables: etchant concentration 110% US. 320/o-37%1 and thickener condition (no thickener, silica thickener, and polymer thickener), although their interactions were not significant. One-way ANOVA revealed significant differences in the mean intertubular demineralization depths based on the six etching agents at p < 0.00001. Duncan’s multiple-range test was used to test the significance of differences between pairs of means and to rank them at p < 0.05. Both 32% phosphoric acid with polymer and aqueous 37% phosphoric acid resulted in the mean deepest demineral- izations, which were statistically similar. The shallowest intertubular demineralization was associated with the use of 10% phosphoric acid with silica, which was not statistically different from 35% phosphoric acid with silica at p < 0.05. The 10% phosphoric acid with polymer and aqueous 10% phospho- ric acid were ranked in the intermediary Duncan’s subset.

pH Measurements The pH was directly measured with a digital pH electrode (WTW 537, Wissenschaftlich Technische We&&&en GMBH, Weilheim, Germany), and double-checked with a microcomputer pH meter (Model P207, Consort, Turnhout, Belgium). Prior to measuring each of the products, the pH electrode was calibrated with a pH 4.00 standard solution 0 for the WTW 537 model, and with both pH 4.00 and pH 7.00 standard solutions for the Consort I'207 model. Three measurements were made and a mean calculated.

Statistical analysis. The data obtained from the measurement of demineralization were subjected to one-way and two-way analyses of variance (ANOVA). A post-hoc Duncan’s test was used to compare and rank specific mean values at a confidence level of 95%. A correlation coefficient between deepest demineralization of intertubular dentin and pH was calculated. The statistical analysis was carried out with the software system (SPSS/PC+, SPSS, Inc., Chicago, IL).

RESULTS The mean intertubular demineralization and standard deviations are listed in Table 2. Two-way ANOVA

264 Perdig3o et al./SEM of acid-etched dentin

The measured pH values are given inTable 1. A significant correlation was found between the deepest intertubular demineralization and pH (r = -0.6663; p < 0.0001).

Representative SEM micrographs are shown in Figs. 1 to 6. Figs. 7 and 8 represent the two different mechanisms 01 interaction of the etchants with dentin. When the tubules were observed in cross section from an occlusal perspective, the specimens of Groups 2 and 5 displayed a residual layer of

I 1 NO Thin Yes 2 NO Thick Yes 3 2.44 i OJb NO Thin No 4 No Thick Yes 5 No Thick Yes

3 1 NO Thick Yes 2 ‘No Thin Yes 3 2.70 i: 0.54b No Thick Yes 4 No Thick Yes 5 Yes Thick Yes

1.60 + 0.14”

Yes Thick Yes Yes Thick Yes Yes Thin Yes Yes Absent Yes Yes Thin No

clustered silica particles attached to the pores between the inter-tubular collagen fibers, although in most of the specimens they did not reach the outer per&&mar collagen (Figs. 2a, b and 5a, b). Fig. 2c shows one tubule in which the silica particles reached the ~peritubular collagen. However, the specimens of Group 5 had fewer residual particles than those of Group 2. These silica particles seemed to enter a few pores on the superficial collagen network (Fig. Zd).

In addition, a cuff of super&al mineralized peritubular dentin was visible at the entrance of each tubule in all the specimens .of Groups :2 and 5 (Figs 2a and 5a). IIowever, the mineral rim was more prominent in Group 2 and had an empty circun-&erentral groove around it (Fig. 2b). ThelcufIcouId alsobe observed in one specin& bf Group 3 (Table 2).

When dentinal tubules wererobserved longitudinally ,fiom a lateral perspe&e,the interaction oftbe acids with dentinal surfaces resulted in a layer of demineralized dentin with exposed collagen. Peritubular dentin was generally more

accessible to the acid through the patent tubule than was intertubular dentin, and it formed a funnel-shaped demineralization pattern (Figs. 4c and 6~). In Group 2, the pattern of peritubular demineralization was completely different, as the vertical depth ofintertubular and peritubular demineralization was similar (Fig. 2c). The depth of peritu-bular deminer- alization varied widely within each of the other groups.

A superficial layer of residual material was observed in all the specimens of Groups 1, 3,4, and 6, but was not readily apparent for the specimens of Groups 2 and 5 (Figs. 2c and 5c), in which granular silica deposits were evident in some areas (Table 2 and representative micrographs by group). In the specimens of Group 4, the entrance of the tubules was wider than in the other groups, and the surface residual layer blocked most of the intertubu- lar pores (Figs. 4a, b, d).

The submicromillimeter hiatus, which is a nano-space at the transi- tion between the zone of packed collagen fibers and the unaffected dentin (Perdigao et al., 1995), could be observed in all the specimens of Groups 3,4, and 6, in four specimens of Groups 1 and 5, and’in two specimens of Group 2, which showed the least frequent hiatus for- mation (Table 2 and representative micrographs by group): When the snecimen was scanned along its Whole diameter, the hiatus appe’aed as a continuous semi&led space, or as an intermittent phenomenon.

DlSCUSSlON

Acid gels may have a 1Wfold lower dBusion coefficient than liquid etchants (Pashley et al., 1992); however the demineralizing effects of phosphoric acid etchants on dentin may be similar, regardless of their viscosity (Ishikawa et al., 1989). Similarly the enamel bond strengths obtained with enamel etched with’,three different viscosities of phosphoric acid are not statistica‘lly Merent (Guba et al., 1994).

The deepest demineralization of intertubular dentin values may be under-estimated. First, it has been demonstrated that most acids demineralize dentin by irreversibly removing part of the substrate, iyhich cannot be detected unless a varnish is used to protect part of the dentm surface (Van Nieerbeek ed all, .1992; Inokoshi et al., 1993). For exan%ple, acids can penetrate 8 pm, although- the depth observable in the corresponding SEM micrograph is4 pm (hiokoshief al., 1993). Second, it has been shown that there is shrinkage of demineralized dentin in every step used in SEM preparation (Carvalho et al., 1995; Perdigao et al., 1995).

Dental Materials/July 1996 265

Fig, 2. Dentin etcheu WI 10% phos- phoric acid Ultra-EtchM (Grou? 2). a) Cross-sectioned tubules with intense intertubular microporosity, residual silica particles, and a cuff of peritubular dentin (arrows). The outer peritubular collagen is observed between the peritubular cuff and the dentin surface (asterisks). Bar = 3 pm. b) Entrance of a tubule showing the peritubular cuff and the lumen (L) of the tubule. The intertubular dentin is less porous around the tubule (black arrow). The silica particles do not enter all the intertubular pores (white arrows). A circumferential groove developed around the peritubular cuff (white asterisk). Bar = 750 urn. c) Lateral view of the transition between demineralized and unaffected dentin. Collagen fibers are embedded in min- eral (asterisks); residual particles are present on the surface (white arrows). A peripheral groove is observed around the cuff (black arrow). S - dentin surface; D-unaffected dentin. Bar ?? 750 urn. d) Intertubular area showing collagen fibers separated by pores. The subsurface collagen network is observed through these pores. The superficial fibers are flattened and appear different from subsurface fibers observed in Fig. 2c. Residual particles are embedded in the collagen web (arrows). Bar ??

500 pm.

Fig. 3. Dentin etched with 10% phos- phoric acid All-Etchrk (Group 3). a) Cross-sectioned tubules showing intense intertubular microporosity, outer peritubular collagen fibers, and fibrous structures inside the tubules (asterisks), without peritubular cuffs. Residual smear layer agglomerates are observed (arrows). Bar = 3 urn. b) Entrance of a tubule showing less porous intertubular dentin (black arrow). Beneath the outer peritubular exposed collagen appears the miner- alized peritubulardentin (whitearrow). Bar ??750 urn. c) Lateral view of the transition between demineralized and unaffected dentin, showing a continu- ous collagen network on the left, and the genesis of a submicromillimeter hiatus on the right. Residual particles are present on the surface (white ar- rows). S - dentin surface. D - unaf- fected dentin, Bar = 1.5 urn. d) Transi- tion between the tubule wall and intertubular dentin. The intertubular dentin was demineralized superfi- cially (black arrow) and laterally (white arrow), via tubules. The triangular- shaped pattern of peritubular dem- ineralization is shown (oval). The direction of the outer peritubular fibers is mostly circular, while the intertubular fibers are randomly oriented.The outer peritubular fibers show the characteristic collagen banding.Asubmicromillimeter hiatus is observed above the unaffected dentin. Bar = 750 pm.

266 Perdigao et al./SEM of acid-etched dentin

Fig. 4. Dentin etched with aqueous 37% phosphoric acid (Group 4). a) Cross-sectioned tubules showing intertubular microporosity, exposed outer peritubular collagen fibers, and a fibrous structure (arrow). Bar = 3 pm. b) View of the entrance of a tubule. The inter- tubular dentin is less porous than in previous correspond- ing micrographs (Figs. lb, 2b, 3b). Bar = 750 pm. c) Lateral view of the transition between demineralized and unaffected dentin with a small hiatus along the whole periphery of the unaffected dentin. A layer of residual material covers the surface (white arrows). The triangular-shaped pattern of peritubular demineralization is also shown (black arrows). A fibrous core is observed in one of the tubules (black asterisk). S - dentin surface; D - unaffected dentin. Bar ??

1.5 pm. d) High magnification of the transition between unaffected and demineralized dentin. PC - outer peritubular collagen; IC - intertubular collagen; H - hiatus. Bar ??

500 vm.

Fig. 5. Dentin etched with 35% phosphoric acid Ultra-EtchTM (Group 5). a) Cross-sectioned tubules showing intertubular microporosity and exposed outer peritubular collagen fibers. A cuff of peritubular dentin is present in most tubules. Residual silica par- ticles are observed (narrow white arrows), along with a lat- eral tubular branch that opens on the surface (wide white ar- row). Bar = 3 pm. b) Entrance of the tubules showing intertu- bular porosity and a cuff of peritubular dentin. Bar = 1.5 pm. c) Transition between demineralized and unaffected dentin showing a submicro- millimeter hiatus (white aster- isk) and collagen fibers. The tri- angular-shaped pattern of peritubular demineralization is shown (blackarrows), although it is not as pronounced as in Groups 3, 4 and 6. S - dentin surface. D- un-affected dentin. Bar = 750 pm. d) Lateral view of a tubule,showing the transi- tion between the unaffected peritubular dentin (white aster- isk) and the demineralized peritubular dentin. Note the characterisitc collagen band- ing in some of the outer peritubular collagen fibers. The fibers of the deepest area are surrounded by mineral crystals (black oval). Bar ??500 pm.

Dental Materials/July 1996 2 67

Fig. 6. Dentin etched with 32% phosphoric acid Uni-EtchTM (Group 6). a) Cross-sectioned tubules showing intertubular microporosity and exposed outer peritubular fibers. In some areas the porosity is blocked (circles). A collar of less porous dentin exists around each tubule (arrows). Bar = 3 pm. b) Entrance of a tubule with porous intertubular dentin and a patent lumen. The subsurface collagen fibers are visible through the pores. Bar = 750 pm. c) Transition between demineralized and unaffected dentin showing the submicromillimeter hiatus (white asterisk) and the triangular-shaped pattern of peritubular demineralization (black arrows). The transition between demineralized and unaffected peritubular dentin is observed in the left tubule (wide black arrows). S -dentin surface. D - unaffected dentin. Bar = 1.5 pm. d) Transition between oeritubular and intertubular dentin with a submicromillimeter hiatus (asterisk). Some collagen fibers are embedded in unaffected dentin, while others are disrupted. Bar = 750 pm.

The depth of intertubular demineralization obtained with the viscous polymer-thickened gels was statistically different (p < 0.05) from the depth obtained with the corresponding silica-thickened gels (Table 2). The addition or subtraction by the manufacturer of specific components in the polymer- thickened etchants may have resulted in a stronger demineralization ability than silica-thickened etchants have in corresponding concentrations.

The silica thickened 35% phosphoric acid gel (Group 5) did not penetrate dentin more readily than did the corresponding low concentration, which is in agreement with previous in. vitro studies (Wang and Hume, 1988). However, the 32% polymer-thickened (Group 6) and the 37% unthickened (Group 4) etchants penetrated more deeply than their 10% counterparts. The specific composition and consistency ofboth silica-thickened etchants (thin gels) may have contributed to this difference (Fig. 7 as compared to Fig. 8). It has been reported that high phosphoric acid concentrations may be associated with the formation of calcium phosphate crystals at the dissolution tiont, blocking further penetration of the acid (Wang and Hume, 19881, and that 85% phosphoric acid may fail to etch enamel, which it does in concentrations f?om 10% to 45% (Retiefj 1975). It remains uncertain whether or not the 35% silica-thickened et&ant (Group 5) was able to form more precipitates at the demineralization fi-ont than the other highly concentrated etchants, thus blocking penetration more completely. Nevertheless, there was no evidence of any

268 Perdk#o et al&EM of acid-etched dentin

precipitate associated with the high acid concentrations other than some superficial silica clustering in Group 5.

The acid used in Group 5 resulted in the deepest mean intertubular dentin demineralization of 1.60 m, while the acid used in Group 6 resulted in a deeper mean demineralization front (4.02 pm). Also, the acids used in Groups 2 and 3 resulted in different deepest mean intertubular dentin demineralization, despite their similar concentrations. Manufacturers use thickeners to improve the handling characteristics of their etching gels, some with a polymer, others with silica particles (Perdiggo et al., 1994a). Some silica-thickened gels have a thin consistency while others have a thick consistency (Gubaet al., 1994). other components must interfere with the demineralizing action, such as surfactants. colorants, and disinfectants.

The shallow depth of in&tubular dentin demineralization and the presence of a cuE of superficial peritubular dentin in all the specimens etched with silica-thickened gels (Groups 2 and 5) may have been caused by a less aggressive effect of the respective etchant (Fig. 7). Consequently this may favor the establishment of high dentin bond strengths (Yamaguchi et al., 1989; Perinka et al., 1992), since more calcium might be available. The cuff of peritubular dentin was more prominent in Group 2. It was surrounded by a peripheral groove, which suggests that peritubular dentin was not only accessible through the patent tubule to the acid, but also from the demineralized intertubular dentin side (Fig. 2b). The etchant

Smear Layer

I

Outer Perltubular Cokgen Flben and Perlphen, Intertubular Collagen Fibers

\ \ Peritubular Cuff

Residual Silica

1

Fig. 7. Demineralization pattern of silica-thickened etchants.

__ __....._.. Pe&bunlar

Mode IA

Smear Layer

I

Outa Pwltubu*r Colkgm Fibam and Prlplwd Intatubutmr Coltqon Fiben

1984; Pashley et al., 1993); 2) an intermedi- ate fibrous layer; and 3) a deeper area displayingasubmicromillimeterhiatuswith some scattered mineral crystals and a few randomly disposed collagen fibers (Fig. 8). This sub-micrometer area separated the middle layer from the tmal&&d dentin. The space within this deepest area may be re- sponsible for marginal leakage. It has re- cently been reported that nanoleakage oc- curs at the base of the hybrid layer due to the presence of nanometer-size spaces beneath the hybrid layer (Sane et al., 1994; 1995). Also, if the adhesive monomers fail to penetrate the submicromillimeter hiatus area, the bonding may be susceptible to long- term degradation. Although there is the danger that adhesive resins will not penetrate into the matrix as deeply as acidic conditioners (Pashleyet al., 1992;Van Meerbeeket al., 19921, Raman spectroscopic analyses have shown that resin penetrates the entire demineralization depth, regard- less of the concentration of the et&ant Wan Meerbeek et al., 1993b).

A. ne&bilized hW*n ---..L

The presence of the submicromillimeter hiatuscanbeexplainedbyfourmechanisms,

Fig. 8. Demineralization pattern of polymer-thickened and unthickened etchants.

might have had additional access to the peritubular area by a lateral pathway although the acids were described as penetrat- ing preferentially through the tubules (Selvig, 1968). Nevertheless, recent research, using dilute acids, has demonstrated that the rate of acid penetration is similar in bothintertubular and peritubular dentin (Marshallet al., 1995). In some areas, the groove seemed to communicate with an unfilled area (Fig. 2b). This is unlike the effects of the unthickened etchants (Groups 1 and 4) and the etchants thickened with polymer (Groups 3 and 6) which followed a more conventional perimbular demineralization pattern Selvig, 1968) (Fig. 8), previously described as modes I and II (van Meerbeeket al., 1992).

Higher magnifications of the demineralized dentin zone partially confirmed previous TEM findings (van Meerbeek et al., 1993a) and Field Emission SEM (Perdigao et al., 1995). Under Field Emission SEM, three layers could be distinguish& 1) a super&&l layer with residual material, which may be a residual smear layer and/or denatured collagen (Bowen et al.,

two of which possibly involve acid kinetics: 1) Carbonate is converted to carbon dioxide, and both calcium and phosphate ions are liberated during acid-etching of dentin (Pashley et al., 1992) and may be liberated faster than they can diffuse from the site. The accumulation of these reaction prcducts may limit further penetration of protons from the acid (Pashley et al., 1992). While the etching agent is being rinsed oe these reaction products may be washed away, leaving a semi-hlled space at the bottom of the demineralized dentin layer, which could be filled with the unfilled resin, regardless of the depth (Van Meerbeek et al., 199313); 2) Etching agents are hypertonic: the

osmolality of 37% phosphoric acid is 6070 mOsm/kg compared to isotonic saline at 290 mOsm/kg (Pashley et al., 1992). Application of phosphoric acid could osmotically draw fluid horn the dentin toward the surface (Pashleyet al., 19921, which could lift the collagen away Tom the unafliected dentin, resulting in an open space. Ideally one should use isotonic acids as dentin et&ants, which have been reported to be well tolerated both by cell cultures and by monkey pulps (Mjor et al., 1982); 3) Carbon dioxide gas entrapped within the polymer-thickened gel may lift the collagen upward and 4) The fracture of the dentin disk could cause displacement of the collagen network toward the application of the force. However, the submicromillimeter hiatus was not observed in five of the 30 specimens.

Collagen banding was present in the peritubular region in one of each specimen of Groups 1,3,4, and 6, in two specimens of Group 2 and three specimens of Group 5. The lack of collagen banding may be caused by the demineralization effect of acids; after demineralization, dentin collagen is in a

Dental Matervh!dJJuly 7996 269

destabilized, but not denatured, state susceptible to enzymatic degradation (Scott and Leaver, 1974; Okamotoet al., 1991). In fact, it has been reported that the presence of hydroxyapatite crystals may stabilize collagen and thus prevent its denaturation and collapse (PashIey, 1992; Sano et al., 1994). The non-collapsed appearance of the intertubular collagen web after application of the acids may have been caused by the presence of residual hydroxyapatite crystals within the highly intertwined network of fibers (Prockop et al., 1979). It has been shown that residual crystals remain intimately associated with the collagen fibers &r demineralization, while interl?briJlar crystals are preferentially dissolved (Selvig, 1968). Some positive ions, such as the calcium, ferric, and aluminum ions in some etchants, might also stabilize collagen and decrease the depth of dentin demineralization (Wang et al., 1991; Nakabayashi, 1992; PashIey et al., 1992; Sano et al.,

1994). Cupric ions included in the primer have been reported to promote the interfacial initiation of polymerization and to increase the tensile dentin bond strengths between dentin etched with 10% phosphoric acid and a 4-META-based resin (Imai and Ikemura, 1994). SimiIarlx the silica contained in some acid gels might interact with the decalcified dentin surface and near-surface as a result of the formation of a siIicc+phosphate layer during etching of dentin (Ruse and Smith, 1991; Eliades, 1993). This interaction might result in similar depths of demineralization for different concentrations of silica-thickened phosphoric acid gels (Perdiggo and &v&1994), as observed for Groups 2 and 5. The silica was not removed even aRer vigorous rinsing, since it may have formed bridges with residual calcium ions (Wilson and Nicholson, 1993) that remained in the decalcified dentin area Oran Meerbeek et aZ., 1993a) and bonded to negatively charged organic groups on the dentin surface.

The measured pH of the six etchants ranged from -0.43 to 1.31. The etchant used in Group 2 displayed the highest measured pH (1.3 11, which may have been responsible for its shallow dentin penetration. Sin&& the measured pH of the etchant used in Group 5 was the highest among high- concentrated etchants, and its penetration into dentin was statistically lower than the etchants used in Group 4 and in Group 6. There was a significant correlation between depth of etching and pH at p < 0.0001.

The hypothesis advanced is rejected. The results obtained suggest that similar concentrations of phosphoric acid etchants containing distinct thickeners result in different demineralization depths as weIl as different morphologies of etched dentin. Other factors may affect the mechanism of action of the phosphoric acid-based etchants used in dentistry

Further studies should combine isotonic solutions of organic and inorganic acids with different thickeners. The interaction of different acid concentrations with both dentin and enamel should be evaluated, since low-concentration etchants are being used increasingly, and perhaps ill-advisedly, in dental adhesion (SwiR and Cloe, 1993; Trio10 et al., 1993).

ACKNOWLEDGMENTS Based on a chapter of a thesis submitted to the graduate faculty, Catholic University of Leuven (KUL) by J. Perdigao, in partial fuI6Ihnent of the requirements for the Ph.D. degree.

We thank Prof J. Vieira, Department of Ceramics and Glass, The University of Aveiro, Portugal, for extensive instrumental support. This research was supported in part by research grants

from the Calouste Gulbenkian Foundation, Portugal, from the National Fund of Scientific Research, Belgium, from Ultradent Products Inc., USA, and from Bisco Dental Products, USA.

Received January 12, 1996 / Accepted June 24, 1996

Address correspondence and reprint requests to:

<Jorge Perdig%

R. Luis F Sommer, 86,l Esq.

2330 Entroncamento PORTUGAL

Phone +351-49-719022

Fax 1 +351-49-716212

REFERENCES

Bowen RL, Eick JD, Henderson DA, Anderson DW (1984). Smear layer: Removal and bonding considerations. Oper Dent Suppl. 3:30-34.

Boyde A, Wood C (1969). Preparation of animal tissues for surface-scanning electron microscopy JMicrosc 90:221-249.

Buonocore MG (1955). A simple method of increasing the adhesion of acrylic tiIIing materials to enamel surfaces. J Dent Res 34849-853.

Car&ho R,Yoshiyama M,Tagami J, Sano H, Brewer P, Pashley D (1995). Shrinkage of demineralized dentin during preparation for SEM. J Dent Res 75:39O,Abstr. No. 2981.

Chiba M, Itoh K, Wakumoto S (1989). Effect of dentin cleansers on the bonding efficacy of dentin adhesive. Dent Mater J 8:76-85.

Delong A (1993). Electron sources for electron microscopes. European Microscopy &Analysis 2619-11.

Eick JD (1992). Smear layer - Materials surface. Proc Finn Dent Sot 88:225-242.

EIiades GC (1993). Dentine bonding systems. In: Vanberle G, Degrange M, WiIIems G, editors. State of the Art of Direct Posterior FilIing Materials and Dentine Bonding. Leuven: Van der Poorten, 49-74.

Fusayama T, Nakamura M, Kurosaki N, Iwaku M (1979). Non-pressure adhesion of a new adhesive restorative resin. J Dent Res 58: 1364-1370.

Goldstein JI, Newbury DE, Echlin P, Joy DC, Fiori C, Lifshin E (1981). Scanning electron microscopy and x-ray microanalysis -A text for biologists, material scientists, and geologists. New York: Plenum Press, 19-51.

Guba CJ, Co&ran MA, Swartz ML (1994). The effects of varied etching time and etching solution viscosity on bond strength and enamel morphology Oper Dent 19:146-153.

Gwinnett AJ, Matsui A (1967). A study of enamel adhesives. The physical relationhip between enamel and adhesive. Arch Oral Bid 12:1615-1620.

Imai Y, Ikemura R (1994). Influence of primers containing cupric ion on bonding of dentin treated with phosphoric acid. Dent Mater J 13:66-75.

Inokoshi S, Hosoda H, Harnirattisai C, Shimada Y (1993). Interfacial structure between dentin and seven dentin bonding systems revealed using argon ion beam etching. Oper Dent 18%16.

Ishikawa K, Ito S, HataY ( 1989). Permeability of etching agent constituents through dentin. Dent Mater J 8:164-174.

Kubo S, Finger W,J, MuIler M, Podszun W (1991). Principles and mechanisms of bonding with dentin adhesive materials. J Esthet Dent 3:62-69.

270 Perdk_xTo et aMSEA of acid-etched dentin

Marshall GW, BaIooch M, Kinney JH, MarshalI SJ (1995). Atomic force microscopy of conditioning agents on dentin. J Biomed Mater Res 29:1381-1387.

Mjiir IA, Hen&en-Pet&men A, Bowen RI, (1982). Biological assessments of experimental cavity cleansers: Correlation between in vitro and in vivo studies. J Dent Re.s 61:967- 972.

Nakabayasbi N ( 1992). Adhesive bonding with CMETA. Oper Dent Suppl. 5:125-130.

Nakabayashi N, Kojima K, Masuhara E (1982). The promotion of adhesion by the infiItration of monomers into tooth substrates. J Biom.ed Mater Res 16:265-273.

OkamotoY, Heeley JD, Dogon IL, Shintani H (1991). Effects of phosphoric acid and tannic acid on dentine collagen. JOral Rehabil18:507-512.

Pashley DH (1992). The effects of acid-etching on the puIpodentin complex. Oper Dent 17:229-242.

PashIey DH, Tao L, Boyd L, King GE, Horner JA (1988). Scanning electron microscopy of the substructure of smear layers in human dentine. Arch Oral Biol33:265-270.

PashIey DH, Horner JA, Brewer PD (1992). Interactions of conditioners on the dentin surface. Oper Dent Suppl. 5:137- 150.

Pashley DH, Ciucchi B, Sano H, Horner JA (1993). PermeabiIity of dentin to adhesive agents. Quintessence Int 24618-631.

Perdigao J, Denehy GE, Swift EJ (1994a). Silica contamination of etched dentin and enamel surfaces: A scanning electron microscopic and bond strength study Quintessence Int 25:327-333.

Perdiggo J, Lambrechts P, Vanherle G (1994b). A morphological Field Emission SEM study of acid-etched dentin. J Dent Res 74916, Abstr. No. 42.

Perdiggo J, Swift EJ (1994). Adhesion of a total-etch phosphate ester bonding agent. Am JDent 7:149-152.

Perdiggo J, Lambrechts P, Van Meerbeek B,Vanherle G, Lopes ALB (1995). A field emission SEM comparison of four post- &&ion drying techniques for human dentin. J Biomed MaterRes 29:1111-1120.

Per&a L, Sano H, Hosoda H (1992). Dentin thickness, hardness and Ca-concentration vs. bond strength’of dentin adhesives. Dent Mater 8:229-233.

Prockop DJ, Kivirikko KI, Tuderman L, Guzman NA (1979). The biosynthesis of collagen and its disorders (fu-st. of two parts). N Engl J Med 301:13-23.

Retief DH (1975). The use of 50 per cent phosphoric acid as an etching agent in orthodontics: A rational approach. Am J Orthodont 68:165-178.

Ruse ND, Smith DC (1991). Adhesion to bovine dentin - Surface characterization. J Dent Res 70:1002-1008.

Sano H, Shono T Takatsu T, Hosoda H (1994). Microporous dentin zone beneath resin-impregnated layer. Oper Dent 19:59&l.

Sano H, Takatsu T, Ciucchi B, Horner JA, Matthews WG, PashIey DH (1995). Nanoleakage: Leakage within the hybrid layer Oper Dent 20118-25.

Scott PG, Leaver AG (1974). The degradation of human collagen by trypsin. Connect Tissw Res 2:299-307.

Selvig KA (1968). Ultrastructural changes in human dentine exposed to a weak acid. Arch Oral Biol13:719-734.

Swift EJ, Cloe BC ( 1993). Shear bond strengths of new enamel etchants. Am J Dent 6:162-X4.

Trio10 PT, SwiR EJ, Mudgil A, Levine A (1993). Effect of etching time on enamel bond strengths. Am J Dent 6:302- 304.

Van Meerbeek B, Inokoshi S, Braem M, Lambrechts P,Vanherle G (1992). Morphological aspects of the resin-dentin inter- diffusion zone with different dentin adhesive systems. J Dent Res 71:1530-1540.

Van Meerbeek B, Dhem A, Goret-Nicaise M, Braem M, Lambrechts P,Vanherle G (1993a). Comparative SEM and TEM examination of the ultrastructure of the resin-dentin interdiffusion zone. J Dent Res 72:495-501.

Van Meerbeek B, Mohrbacher H, CeIis Jp, Riots JR, Braem M, Lambrechts P,Vanherle G (1993b). Chemical characteriza- tion of the resin-dentin interface by micro-Raman spectros- copy J Dent Res 72~1423-1428.

Wang J-D, Hume WR (1988). Diffusion of hydrogen ion and hydroxyl ion fiorn various sources through dentine. Int Endodont J 21117-26.

Wang T, Nikando T, Nakabayashi N (1991). Photocure bonding agent containing phosphoric methacrylate. Dent Mater 7:59-62.

Wilson AD, Nicholson JW (1993). Acid-base Cements. Their Biomedical and Industrial Applications. Cambridge: Cambridge University Press, 119-122.

Y&maguchi R, Powers JM, Dennison JB (1989). Parameters affecting in vitro bond strength of composites to enamel and dentin. Dent Mater 5:153-156.

DenhI MaterialdJJuly 1996 271