Studies on the Adhesion of Glass-ionomer Cements to Dentin A. LIN, N.S. McINTYRE, and R.D. DAVIDSON Surface Science Western, Room 1, Western Science Centre, The University of Western Ontario, London, Ontario, Canada N6A 5B7 This study investigated the bonding mechanisms of glass-ionomer cement to dentin. The approaches included mechanical determination of bond strengths, analysis of surface morphology by means of scanning electron microscopy (SEM) and confocal microscopy, and measurement of chemical changes of fracture bond sites by means of x-ray photoelectron spectroscopy (XPS) and secondary ion mass spectrometry (SIMS). The highest bond strengths were obtained with light-cured glass-ionomer cement. SEM and confocal images showed evidence of mechanical interlocking of cement in dentinal tubules. SIMS depth profiles confirmed the ion-exchange process between the light-cured glass-ionomer cement and the dentin surface. From corresponding XPS results, it was clear that the adhesion characteristics were significantly affected by light-curing and the chemical structure of the polymer. J Dent Res 71(11):1836-1841, November, 1992
Introduction. One of the most important aspects of restorative dentistry is the attachment of aesthetic replicas to the remaining part of the dentition. Marginal apposition between the filling materials and walls ofthe prepared cavity is essential ifleakage oforal fluids is not to result in recurrent caries. The attachment ofthe materials to the tissue surface is therefore of the greatest significance in many dental procedures and is often the most important factor in controlling clinical success. Adhesion can be defined as the attraction exhibited between the molecules of different materials at their interface. The criteria for adhesion to be achieved include (a) a clean substrate for intimate access of adhesive to its surface, (b) complete wetting of the substrate surface, and (c) liquid-to-solid transformation of the adhesive. Toxicity, moisture, temperature, and the chemical nature of the substrates present further restrictions which need to be dealt with in the design of an adhesive component. Chemical variation of the substrate, differences in the coefficient of thermal expansion, and volume and wetting changes on solidification will all affect the final performance of the adhesive. In recent years, glass-ionomer cements have been of interest in the area of dental materials. The major advance in the glassionomer cement is its ability to adsorb permanently to the hydrophilic surfaces of hard oral tissues, thus offering the possibility of sealing margins developed at the tissue interfaces during restorative procedures. The composition ofthe glass-ionomer cements is complex and varied. The basic component ofthe glass is a calcium aluminosilicate containing some fluoride. The acid is a polyelectrolyte, which is a homopolymer or copolymer of unsaturated carboxylic acids. The glass-ionomer cement sets as a result of a reaction between an acid and a base, with the product of the reaction, a hydrogel salt, acting as a binding matrix. The development and application ofthe system have been thoroughly discussed by Wilson and McLean (1988). More recently, light-curable systems have become available commercially. These materials have dual-curing Received for publication December 20, 1991 Accepted for publication June 6, 1992 This work was funded by 3M Canada Inc. and the University Research Incentive Fund of the Government of Ontario.
capabilities, i.e., the acid-base reaction is initiated on mixing, but proceeds slowly and is accelerated by a light-activated polymerization mechanism. Many hypotheses for the actual bonding mechanisms of glassionomer cements have been introduced. There is a consensus that the mechanism for bonding to enamel is almost entirely a result of ionic and polar forces (Wilson and McLean, 1988; Smith, 1989). More complex interactions are involved when bonding to dentin surfaces is considered. Bond strengths ofglass-ionomer cements to dentin have been determined by a variety oftest methods and under different conditions of testing. Unfortunately, because ofthis, it is difficult for the results of these studies to be compared. The present work was an effort to study glass-ionomer bonding by means of a number of chemical and physical techniques. The approaches used in this study included: the mechanical determination of bond strengths, analysis of surface morphology by means of scanning electron microscopy (SEM) and confocal microscopy, and measurement of chemical changes of fracture bond sites by means of x-ray photoelectron spectroscopy (XPS) and secondary ion mass spectrometry (SIMS). Confocal microscopy can be used to obtain thin optical sections lan
16-
Fj Immediate
E 14
12 -
* 24 hours to enamel
T
T
.-
IL co
T
10 -
a.
C
(I
8
0
7-
6
4.
2
F1ir.
0-
LCGI
LCGI/chem
Type of Glass
LCGI/chem/ight
lonomer
GIC
Cement
Fig. 1-Results of bond strength tests for light-cured glass-ionomer cement (LCGI), light-cured glass-ionomer cement which was chemicallycured (LCGI/CHEM), and conventional glass-ionomer cement (GIC), showing means and standard deviations. Tests were run immediately and after 24 h to dentin, and after 24 h to enamel.
Downloaded from jdr.sagepub.com at Bobst Library, New York University on April 20, 2015 For personal use only. No other uses without permission.
1836
to dentin
24 hours to dentin
ADHESJOY STUDIES N
Vol. 71 No. 11
Fig. 2-cananingele£ctro micrgph of iled
a.
Sr gp ffiled bond f area on dntirn eltn miaph Fg. 3Smng
below te surfers of the specimens, without te roal problems ofth onsection ir transmid lgit m oscop or tie drng of biolocal samples r tLhe SEML WatsoV n (1989) has bund tis tchnque to be usf h determinng he isiution of adhesives withi he toot and overig restoration. Xra photoelect o spetrosc py is a tchnqu whiv h gives quatitative inormat on virtuy l elements (except hydrogen and heliu) on or within a 3-nm-thk surface ayer. Te detction limit is noaily 0.I at% In addition iis often possiblefbr se information to e obtamd about the chemical enviromen in vhich these elements rside. SIMS extends the surice serstiity ranrg provided by XPS The eleme italdetection limit often extends to0d ppin (aomic) and below. Areas as smial as 50 pm in diameter can be analyzed. Spectra are genmratcd b a primary ion beam whic remoyes the analyte fro the surface by a sputtering processc This process ystematically erodes the outer surface and abyows a concentraton profi c be measured as a inion of depth in the surfie. Sueh profiles generate much information eoacernina trace and major elemental concentration gradients in he outer surface. his technique could thertore be used or inasurtmnat of chan s in the -
ASS-IONOME CEMENTS
emicaly-cured LCGI.
1837
ea
n din
fr
concentration o elem cents a ad neai the iterce between the
gass-ionomer cement and detin. Materials and methods. Bond stre tss.Dent and enamel samples were prepared fbr bond strength tests Ten samples were prepared for each test Flat dentin rnd enamel surf ces Pom freshly extracted bovine teeth wr prepared by wet-grindng on a series o silicon carbide papers from 120- to 600t. The specimens were then washed copiously wth distilled water and mounted in poly(methyl methacrylate) resinmQuikmnourt) in a sanless steel ning The tooth surface was ar-ded. The lass-ionomer cement was prepared according to manufacturers instrucions, and a hin layer (0.1 mm) was placed r the cenrtr of te e exposed area with perforated acetate sheet ven thickness would be ensured. The glass used so at ionomers ested include conventional glass-ionomer cement (Glass lr omer Liner, 3M St. Paul MN) and a light-curable glass-ionomer cemer t Vitrebonid 3M). T powder component of both cements was coriposed primarily of a radiopaque fluoroaluminosilicate glass powder The liquid solution for he conventional liner was composed of a poly(a ryli acid) terpolymer. The light-curable cement had a
Fig 4-Coafocal scanamag microscope image showing anterface between IFg 5-Coafbcal scanning microscope image showing interface between LCGI and dentin A ea I, dentin; Area 2, fluoresceatoeIabeled CGI; and chemically red LCGI and derin aAre 1 dentin, Area 2 fluorescence Are 3 fluorecenel4abeledDownloaded adhesive.from jdr.sagepub.com at Bobst Library, New York University onlabeled LCGI; and Area 3, fluorescence-labeled adhesive April 20, 2015 For personal use only. No other uses without permission.
1838 Li~~~~~~~~~~~~etal. LIN 1838 et al. ~~~~~~J Dent Res November 1992 50000
-.1
I'll
1-11.
ll ......
1.1-
Flood Gun 1.0 eV 45000 0 Is 40000
Pis~~~~~~~~~
35000 30000
rM
10 m
0
25000 F-
0 Q
C-)
20000 ~-
15000 F F is
10000
N J n C
Zn 2p3
Ca
5000 600
500
400
1100
Binding Energy (nV)
Si
Al
1000
900
800
700
600
500
400
300
200
100
Binding Energy (eV)
Fig. 6-X-ray photoelectron spectrum of polished dentin surface
Fig. 7-X-ray photoelectron spectrum of fracture site on dentin after bonding with LCGI.
liquid that was an aqueous solution of a poly(acrylic acid) copolymer containing pendant methacrylate groups together with hydroxy(methyl methacrylate) and photoinitiator. The ionomer either was light-cured with a 470-nm blue light (Visilux, 3M) or was allowed to cure chemically. An adhesive (Scotchbond 2, 3M) was applied to the surface ofthe ionomer and light-cured. A Teflon mold was filled with a composite material (P-5O, 3M) and centered over the ionomer layer. The composite was also light-cured. The mold was kept clamped in place for 15 min at room temperature. The bonded assembly was then removed from the clamps and stored in a humidity oven at 100% relative humidity at 370C for 24 h until tested. Those that were tested for immediate strength were not stored in the humidity oven. After 24 h, the specimens were transferred to a testing holder which could be attached to the moving cross-head of an Instron testing machine. A wire was attached to the load cell and used to pull the buttons off the tooth surface. The tensile load required to cause fracture of the bonds was recorded at a cross-head speed of 2 mm/mmn. The bond strength was calculated from the load required to cause fracture ofthe bond divided by the area ofthe exposed tooth
500UU L
surface.
Scanning electron microscopy (SEM).-An ISI Model DS-130 scanning electron microscope was used for examination of the surface dentin and enamel structure before bonding and also the interfacial structure between ionomer and dentin. The correspond-
hydrocarbon
(a)
Flood Gun 1.0 eV
hydroxyl
carbonyl /' carboxyl
(b)
I
45000
P
........4 1-C.
-
0 1s
carbide
40000
C is 35000
30000
0) r
3
25000
0
0
(c)l 20000
15000
10000
Nooa 5000
Si 2p .I.
.I.
I.
.I.
.I.
..
Binding Energy (eV)
Fig. 8-X-ray photoelectron spectrum of fracture bonding with chemically-cured LOGI.
295.0
BINDING ENERGY (eV)
275.0
....
site
on
dentin after
Fig. 9-Carbon is spectra for (a) dentin, (b) LCGI bond area, and (c) chemically-cured LCGI bond area. The moieties responsible for the various peaks are indicated in (a) and (b) and are shown as dotted lines.
Downloaded from jdr.sagepub.com at Bobst Library, New York University on April 20, 2015 For personal use only. No other uses without permission.
11 ADHESION STUDIES ON GLASS-IONOMER CEMENTS V61.Vol 7171 No.No.11
1839 13
DON
16
A
000000000000000
O
000* 0
0 00000og'
6
N
c F Al Si
0 A p
000a0o
4*
#
C4
v
Z"
z
z
9J
A
K~~~~~~~~~~~~~~~. -
05
DET (JmT
Fig. 10--SIMS depth profile through light-cured glass-ionomer cement and into dentin surface.
ing faied surfaces were examined on both the dentin and the glassionomner sides. This was carried out on half of the samples used in the ahoye bond strength tests. Prior to analysis, all sample surfaces were air-dried and sputter-gold-coated. Xlray phtotoeicctro s pcctroscopy XPS .-XPS analyses were carried out obr detection of chemical changes or the surfaces of the other half of the samples flowing bond strength tests. With an Isomet slow-speed saw, samples were cut to a size of 2 mm x 2 mm. Immediately pnior to analysis by XPS, the samples were evacuated in a vacuum chamber (5 x 10- Pa) so that the water vapor pressure would be reduced to alloy for sample introduction into the UH-V spectrometer analysis chamber. The spectrometer used was a customized SSL SSX-100 x-ray photoelectron spectrometer, which uses a monochromatized Al Ku x-ray gun to excite photoelectrons fromr a spot size assmall as 150 gim in diameter. It is possible for the position of the spot wV ere analysis is to take place to be adjusted with aprecisionr of~t 10 gm. Control specimens ofdentin were always analyzed so that the effect of dehydration of samples in the vacuum chamber of the spectrometer could be monitored. For each sample analyzed, broaascans were accumulated for detection of anl elew ments. During these analyses, a spot size of 600 jim and a pass energy of 148. 0 eV were used. For some ofthe peaks, high-resolution spectra of these particular lines (mainly carbon, oxygen, and nitrogen) were obtained so that chemical effects on photoelectron peak positions could be assessed. For these analyses of the core level spectiad a 150 jim spot size and a pass energy of 50.0 eV were used. The charging effect observed during the photoelectron analysis of non-conducting surfaces was controlled by utilization of an electron flood gun and a metal grid placed close to the surface. Secondary ion mass spcctronmetry (SIMS).In order for the interface between the light-cured glass-ionomer cement and the dentin surface to be analyzed, SIMS depth profiles were camrid out through these layers in a prepared sample. Depth profiles were carried out on light-cured glass-ionomer cement, polished dentin, and interfaces between the two substrates. A Cameca IMS-3f secondary ion mass spectrometer was used. An 0- primary beam provided a stable charge condition on the non-conductive samples. Positive s.econdary jons~were monitored, which provided good ion yields for the elements ofinterest. The elements monitored were H1, 12C, i9F, 2WA, 28Si, 31P, "Ca, and 4Zn. Laser scanning confcal microscopy Confocal microscopy was used for the obtaining of thin optical sections below the surfaces of the specimens. This technique provides a unique opportunity for images of the penetration and thickness of the cement and adhesive in the dental restoration to be obtained. Freshly extracted third molar teeth were used for the restorations. Class V cavities were prepared as in normal clinical practice. All cavities were 2 mm wide and 2 mm high and were cut with a type 012 diamond bur running wet in an ultra-high-speed handpiece. Enamel surfaces were etched
Fig. 11!Scarnnirg electron mvicrograph offracture surface after boarding with light-cured glass-irnomenr ceet, showirng rei tags that have penetrated dentirnal tubules,
with 37% phosphoric acid gel for I5 s and then rinsed with distilled water for 30 s Light-cured glass-ionomer cement (Vitrebond, 3M) and conventional glass-ionomer cement (Glass lonomer Cement Liner, 3M) were both prepared according to the manufacturer's instructions. A thin layer of ionomner was placed over the dentin and light-cured if necessary. Dentin primer (Scotchprep, 3M) was applied, to any exposed dentin surfaces, with a sponge for 60 s and was agitated so that the surface would be kept wet for the time period. The primer was ther air-dried. The adhesive resin (Scotchbond 2, 3M) was then applied to the glass-ionoirer cement, primed dentin, and etched enamel surfaces. The adhesive was 1 ghtcured for 20 s. The cavities were then filled with a composite material (P-SO, 3M) and light-cured. The restored teemh samples were kept tor 24 h in aistiiled water at 37 C With an Isomet slow-speed saw, samples were sectioned longitudinally through the centers of the restorations, and the diamond blade was lubricated with water. Thus, the interface was exposed for examination with the microscope. Five restorations of each type (light-cured vs. chemically-cured glass-ionomer cements) were prepareed, giving a total of~teii block faces. To aid in the visualization of the penetration of the glassionomer cement and the adhesive, fluorescent labels were incorporated into the materials. Double-labeling by fluorescing isothiocyanate (FITC) and Texas Red (Molecular Probes Inc., Eugene, OR) was used. The FITC is maximally excited by the 488-nm (hlue) laser line and emits at wavelengths3 to~vhich the. photonmultipliers are highly sensitive. Texas Red is maximally excited at the 514 nm (green) laser line, thus producing perfect, registration for two-channel confocal imaging. FITC was used for labeling the Scotchbond 2 adhesive resin, and Texas Red was used for labeling the polymer portion of the glass-ionomer cement. This produced the best double-labeled images. Scanning images were obtained in a Bio-Rad MRCC 600 laser scanning confocal microscope (Bio-Rad Inc., Camabridge, MA). The scanned samples were examined for gap formations, penetration of materials into the tooth surfaces, and distribution ofthe two materials relative to each other. Images were recorded with a 35-mm camera with T-MAX 100 black-and-white high contrast film (Kodak).
Results. The bond strength results are presented in Fig. 1. The glassionomer cement samples which were light-cured had the highest
Downloaded from jdr.sagepub.com at Bobst Library, New York University on April 20, 2015 For personal use only. No other uses without permission.
1840
LIN et al.
J Dent Res November 1992
bond strength to dentin, both immediately after curing and following 24 h in ambient conditions (p < 0.001). When the same glassionomer cement was only chemically-cured, the bond strength to dentin was lower than that after light-curing, both immediately and after 24 h (p < 0.001). When the glass-ionomer cement was first allowed to cure chemically for 15 min and was subsequently lightcured, 24-hour bond strengths to dentin began to approach the strength of the light-cured glass ionomer to dentin. An SEM examination of the fracture site suggested a bonding mechanism and the nature of bond failure. For the light-cured ionomer, the bonded area showed the presence of a thin layer ofglass ionomer on the surface of the dentin, as seen in Fig. 2. The lightcured ionomer samples which were chemically-cured resulted in a bonded area with very little residual glass ionomer (Fig. 3). However, a smear layer ofdentin debris was observed on the bonded area ofthe glass-ionomer side, showingthat this system failed cohesively within the smear layer. Scanned images of the sections taken with the confocal microscope showed close adaptation of the light-cured glass-ionomer cement to the dentin surface. Penetration of the polymer into the dentinal tubules is clearly seen in Fig. 4. The chemically-cured glass-ionomer cement exhibited less effective penetration into the dentin surface (Fig. 5). The labeled cement had not moved down along the dentinal tubules to the same extent as had the light-cured material. By the sections being double-labeled, the distribution of both the adhesive and ionomer components of the restoration system could be observed. No adhesive penetrated the ionomer layer and smear layer to enter the tubules, except where ionomer
intensity in the control ionomer when compared with the ionomer close to the interface. Near the interface, the glass ionomer had increased amounts of carbon, hydrogen, fluorine, phosphorus, and zinc. SIMS depth profiling confirmed the ionic exchange process between the light-cured glass-ionomer cement and the dentin surface. There was evidence for the movement of ions from the glassionomer cement 1.5 gim into the dentin surface and the movement of calcium and phosphorus ions from dentin 1.0 pm into the ionomer.
Discussion. When glass ionomers come in contact with a dentin surface, the polymer is believed to bond chemically to the surface with the
production of strong adhesive bonds between the two surfaces. When any type of tangential motion is attempted, shear that occurs in the polymer indicates that the interfacial bond is stronger than the cohesive bond in the polymeric material. The bonds in the cohesively-weaker ofthe two materials will fracture preferentially, with the adhesive bond at the interface being stronger than forces in either of the two materials in contact. When sliding is initiated between the surfaces for the light-cured glass ionomers in contact with the dentin or enamel, shear takes place in the ionomer, and an ionomer film is left on the solid surface. In other words, the interfacial bond formed between the ionomer and the solid substrate is stronger than the cohesive bond in the polymer, and therefore, shear occurs in the glass ionomer. When the material was chemically-cured, the glass-ionomer cement fractured at the smear layer. Failure occurred at the weaker adhesive bonds which were was not present. formed at this ionomer/smear layer interface. The XPS results are presented in Figs. 6, 7, and 8. Those dentin Additional stability of the glass-ionomer surface bond is prosurfaces with smear layer primarily consisted of carbon, oxygen, vided by mechanical interlocking of polymer in dentin. Mechanical nitrogen, calcium, and phosphorus. The dentin side of the failed adhesion would naturally be enhanced by greater contact with the bond area for light-cured glass-ionomer cement to dentin appears to irregularities or pores of the substrates. The viscosity of the attract additional silicon, aluminum, fluoride, and zinc ions, pre- materials is therefore important with respect to this particular sumably from the glass-powder component. Calcium and phospho- mechanism. This type ofmechanical interlocking was evident from rus concentrations are also reduced in this area. The dentin side of SEM observations ofthe bond fracture surfaces and scanned confothe failed bond area for chemically-cured glass-ionomer cement cal microscope images of the interface between the ionomer and which was chemically-cured showed mostly carbon and oxygen and dentin. Fig. 11 is amicrographofthe failed bonded surface ofdentin an even smaller amount of calcium and phosphorus. However, there after fracture. Many tubules are occluded with the light-cured glass was no indication of an increased amount of glass particles, since ionomer, and ionomer resin tags protrude from this surface. The only a small amount of silicon was detected at this surface. restoration section images obtained by the confocal microscope High-resolution spectra for carbon are shown in Fig. 9 and clearly showed penetration ofthe light-cured glass-ionomer cement compared for surfaces of dentin and the failed bond areas of dentin through the smear layer and into the dentinal tubules. for light-cured and chemically-cured glass-ionomer cement. When Rates of adsorption on the calcified tissue surface and crossthe chemical states of carbon from the C(ls) spectra for the dentin linking in bulk regions of the material will be different for the lightsurface, light-cured glass-ionomer bond fracture site, and chemi- and chemically-cured glass-ionomer cements. The ionized groups cally-cured glass-ionomer bond-fracture site were compared, car- compete for sites on the surface and for available cations for crossbon contribution from different functional groups could be detected linking during setting. The hydrolytic stability of the sonically at each surface. Singly-bonded C-OH or C-O-C groups were ob- cross-linked network formed after the material has set, and its served on the surface. Polar groups bonded to carbon were expected, adhesive strength to the hydrophilic dentin surface is the result of due to the nature of the wet dentin surface. Carbonyl groups are a balance ofelectrostatic forces (repulsive and attractive) duringthe present on the surface of dentin and are due to the collagen content course ofthe polyelectrolyte's ionization through an acid/base reacin the dentin smear layer. A higher amount of carboxyl groups was tion. For example, rapid ionization once the cation source and the detected on the chemically-cured-ionomer failed-bond site com- polymer are brought into contact might lead to the adsorption of a pared with that on the light-cured ionomer. This indicates a low thin, strongly bound layer of polymer material on dentin. This layer amount of ionic bonds present where the bond failed for the light- does not necessarily allow for a molecular continuity between the cured material. Chemical analysis of interfaces showed that differ- tissue interface and the bulkofthe sample (Belton and Stupp, 1988). ent bondingmechanisms may have been responsible forthe bonding Light-curing causes an initial increase in the ionization rate due to of the light-cured and chemically-cured systems. the production of acid (via the photoinitiator), and this results in a SIMS depth profiles are best compared by ionomer and dentin very strong adsorbed layer on dentin. At the same time, internal interfacial profiles being overlaid with the control profiles of glass- cross-linking reactions result from free-radical-polymerization proionomer cement and dentin, as in Fig. 10. On the control dentin cesses. Thus, the material itselfwill have secondary bonding in the surface, calcium, phosphorus, and hydrogen were of intensities bulk among molecular or between side groups which can then similar to those in dentin near the interface. Carbon, aluminum, prevent the cations from approaching the carboxyl groups for crosssilicon, and fluorine were sharply higher at the surface ofthe dentin linking. near the interface. Zinc intensities were moderately higher at the For the chemically-cured glass-ionomer cement, the reduced interface. Aluminum, silicon, fluorine, and zinc were of similar number of carboxylate groups in the polymer due to the presence of Downloaded from jdr.sagepub.com at Bobst Library, New York University on April 20, 2015 For personal use only. No other uses without permission.
Vol. 71 No. 11
ADHESION STUDIES ON GlASS-IONOMER CEMENTS
side groups can decrease the amount of adsorption on the surface of the dentin. From the SEM and correspondingXPS results, it is clear that light-curing and the chemical structures of the polymer used will significantly affect the adhesion characteristics of the glassionomer cement. Further investigation is needed for better understanding ofthe role ofionization rate on adhesion properties and the driving forces for both chemical bonding and mechanical retention to the dentin surfaces. The combined results from the above experiments have provided a more comprehensive view ofthe interfaces between glass-ionomer cements and dentin substrates. Based on these preliminary studies, further investigations on the nature ofchemical bonding at the interface and the role ofmechanical interlocking would be useful for the purpose of optimizing adhesion to the components of tooth structure in an aqueous environment.
Acknowledgments. We gratefully thank Dr. M. Hubert, Ontario Laser and Lightwave
1841
Centre, University of Toronto, for his skilled technical assistance and use of the scanning laser confocal microscope, and Dr. K Nielsen, 3M Canada Inc., for his assistance with the fluorescence labeling. REFERENCES Belton D, Stupp SI (1988). Adsorption of ionizable polymers on ionic surfaces: poly (acrylic acid). Macromolecules 16:1143-1150. Smith DC (1989). In-vitro performance of glass ionomer cements. In: Proceedings ofthe conference oncorrelationbetweenin-vitro and in-vivo performance of dental materials. Trans Acad Dent Mater (Dublin, Ireland). Watson TF (1989). A confocal optical microscope study ofthe morphology of the tooth/restoration interface using Scotchbond 2 dentin adhesive. J Dent Res 68:1124-1131. Wilson AD, McLean .1W (1988). Glass ionomer cements. Chicago (IL): Quintessence Publishing Co. Inc., 13-39.
Downloaded from jdr.sagepub.com at Bobst Library, New York University on April 20, 2015 For personal use only. No other uses without permission.
Introduction. One of the most important aspects of restorative dentistry is the attachment of aesthetic replicas to the remaining part of the dentition. Marginal apposition between the filling materials and walls ofthe prepared cavity is essential ifleakage oforal fluids is not to result in recurrent caries. The attachment ofthe materials to the tissue surface is therefore of the greatest significance in many dental procedures and is often the most important factor in controlling clinical success. Adhesion can be defined as the attraction exhibited between the molecules of different materials at their interface. The criteria for adhesion to be achieved include (a) a clean substrate for intimate access of adhesive to its surface, (b) complete wetting of the substrate surface, and (c) liquid-to-solid transformation of the adhesive. Toxicity, moisture, temperature, and the chemical nature of the substrates present further restrictions which need to be dealt with in the design of an adhesive component. Chemical variation of the substrate, differences in the coefficient of thermal expansion, and volume and wetting changes on solidification will all affect the final performance of the adhesive. In recent years, glass-ionomer cements have been of interest in the area of dental materials. The major advance in the glassionomer cement is its ability to adsorb permanently to the hydrophilic surfaces of hard oral tissues, thus offering the possibility of sealing margins developed at the tissue interfaces during restorative procedures. The composition ofthe glass-ionomer cements is complex and varied. The basic component ofthe glass is a calcium aluminosilicate containing some fluoride. The acid is a polyelectrolyte, which is a homopolymer or copolymer of unsaturated carboxylic acids. The glass-ionomer cement sets as a result of a reaction between an acid and a base, with the product of the reaction, a hydrogel salt, acting as a binding matrix. The development and application ofthe system have been thoroughly discussed by Wilson and McLean (1988). More recently, light-curable systems have become available commercially. These materials have dual-curing Received for publication December 20, 1991 Accepted for publication June 6, 1992 This work was funded by 3M Canada Inc. and the University Research Incentive Fund of the Government of Ontario.
capabilities, i.e., the acid-base reaction is initiated on mixing, but proceeds slowly and is accelerated by a light-activated polymerization mechanism. Many hypotheses for the actual bonding mechanisms of glassionomer cements have been introduced. There is a consensus that the mechanism for bonding to enamel is almost entirely a result of ionic and polar forces (Wilson and McLean, 1988; Smith, 1989). More complex interactions are involved when bonding to dentin surfaces is considered. Bond strengths ofglass-ionomer cements to dentin have been determined by a variety oftest methods and under different conditions of testing. Unfortunately, because ofthis, it is difficult for the results of these studies to be compared. The present work was an effort to study glass-ionomer bonding by means of a number of chemical and physical techniques. The approaches used in this study included: the mechanical determination of bond strengths, analysis of surface morphology by means of scanning electron microscopy (SEM) and confocal microscopy, and measurement of chemical changes of fracture bond sites by means of x-ray photoelectron spectroscopy (XPS) and secondary ion mass spectrometry (SIMS). Confocal microscopy can be used to obtain thin optical sections lan
16-
Fj Immediate
E 14
12 -
* 24 hours to enamel
T
T
.-
IL co
T
10 -
a.
C
(I
8
0
7-
6
4.
2
F1ir.
0-
LCGI
LCGI/chem
Type of Glass
LCGI/chem/ight
lonomer
GIC
Cement
Fig. 1-Results of bond strength tests for light-cured glass-ionomer cement (LCGI), light-cured glass-ionomer cement which was chemicallycured (LCGI/CHEM), and conventional glass-ionomer cement (GIC), showing means and standard deviations. Tests were run immediately and after 24 h to dentin, and after 24 h to enamel.
Downloaded from jdr.sagepub.com at Bobst Library, New York University on April 20, 2015 For personal use only. No other uses without permission.
1836
to dentin
24 hours to dentin
ADHESJOY STUDIES N
Vol. 71 No. 11
Fig. 2-cananingele£ctro micrgph of iled
a.
Sr gp ffiled bond f area on dntirn eltn miaph Fg. 3Smng
below te surfers of the specimens, without te roal problems ofth onsection ir transmid lgit m oscop or tie drng of biolocal samples r tLhe SEML WatsoV n (1989) has bund tis tchnque to be usf h determinng he isiution of adhesives withi he toot and overig restoration. Xra photoelect o spetrosc py is a tchnqu whiv h gives quatitative inormat on virtuy l elements (except hydrogen and heliu) on or within a 3-nm-thk surface ayer. Te detction limit is noaily 0.I at% In addition iis often possiblefbr se information to e obtamd about the chemical enviromen in vhich these elements rside. SIMS extends the surice serstiity ranrg provided by XPS The eleme italdetection limit often extends to0d ppin (aomic) and below. Areas as smial as 50 pm in diameter can be analyzed. Spectra are genmratcd b a primary ion beam whic remoyes the analyte fro the surface by a sputtering processc This process ystematically erodes the outer surface and abyows a concentraton profi c be measured as a inion of depth in the surfie. Sueh profiles generate much information eoacernina trace and major elemental concentration gradients in he outer surface. his technique could thertore be used or inasurtmnat of chan s in the -
ASS-IONOME CEMENTS
emicaly-cured LCGI.
1837
ea
n din
fr
concentration o elem cents a ad neai the iterce between the
gass-ionomer cement and detin. Materials and methods. Bond stre tss.Dent and enamel samples were prepared fbr bond strength tests Ten samples were prepared for each test Flat dentin rnd enamel surf ces Pom freshly extracted bovine teeth wr prepared by wet-grindng on a series o silicon carbide papers from 120- to 600t. The specimens were then washed copiously wth distilled water and mounted in poly(methyl methacrylate) resinmQuikmnourt) in a sanless steel ning The tooth surface was ar-ded. The lass-ionomer cement was prepared according to manufacturers instrucions, and a hin layer (0.1 mm) was placed r the cenrtr of te e exposed area with perforated acetate sheet ven thickness would be ensured. The glass used so at ionomers ested include conventional glass-ionomer cement (Glass lr omer Liner, 3M St. Paul MN) and a light-curable glass-ionomer cemer t Vitrebonid 3M). T powder component of both cements was coriposed primarily of a radiopaque fluoroaluminosilicate glass powder The liquid solution for he conventional liner was composed of a poly(a ryli acid) terpolymer. The light-curable cement had a
Fig 4-Coafocal scanamag microscope image showing anterface between IFg 5-Coafbcal scanning microscope image showing interface between LCGI and dentin A ea I, dentin; Area 2, fluoresceatoeIabeled CGI; and chemically red LCGI and derin aAre 1 dentin, Area 2 fluorescence Are 3 fluorecenel4abeledDownloaded adhesive.from jdr.sagepub.com at Bobst Library, New York University onlabeled LCGI; and Area 3, fluorescence-labeled adhesive April 20, 2015 For personal use only. No other uses without permission.
1838 Li~~~~~~~~~~~~etal. LIN 1838 et al. ~~~~~~J Dent Res November 1992 50000
-.1
I'll
1-11.
ll ......
1.1-
Flood Gun 1.0 eV 45000 0 Is 40000
Pis~~~~~~~~~
35000 30000
rM
10 m
0
25000 F-
0 Q
C-)
20000 ~-
15000 F F is
10000
N J n C
Zn 2p3
Ca
5000 600
500
400
1100
Binding Energy (nV)
Si
Al
1000
900
800
700
600
500
400
300
200
100
Binding Energy (eV)
Fig. 6-X-ray photoelectron spectrum of polished dentin surface
Fig. 7-X-ray photoelectron spectrum of fracture site on dentin after bonding with LCGI.
liquid that was an aqueous solution of a poly(acrylic acid) copolymer containing pendant methacrylate groups together with hydroxy(methyl methacrylate) and photoinitiator. The ionomer either was light-cured with a 470-nm blue light (Visilux, 3M) or was allowed to cure chemically. An adhesive (Scotchbond 2, 3M) was applied to the surface ofthe ionomer and light-cured. A Teflon mold was filled with a composite material (P-5O, 3M) and centered over the ionomer layer. The composite was also light-cured. The mold was kept clamped in place for 15 min at room temperature. The bonded assembly was then removed from the clamps and stored in a humidity oven at 100% relative humidity at 370C for 24 h until tested. Those that were tested for immediate strength were not stored in the humidity oven. After 24 h, the specimens were transferred to a testing holder which could be attached to the moving cross-head of an Instron testing machine. A wire was attached to the load cell and used to pull the buttons off the tooth surface. The tensile load required to cause fracture of the bonds was recorded at a cross-head speed of 2 mm/mmn. The bond strength was calculated from the load required to cause fracture ofthe bond divided by the area ofthe exposed tooth
500UU L
surface.
Scanning electron microscopy (SEM).-An ISI Model DS-130 scanning electron microscope was used for examination of the surface dentin and enamel structure before bonding and also the interfacial structure between ionomer and dentin. The correspond-
hydrocarbon
(a)
Flood Gun 1.0 eV
hydroxyl
carbonyl /' carboxyl
(b)
I
45000
P
........4 1-C.
-
0 1s
carbide
40000
C is 35000
30000
0) r
3
25000
0
0
(c)l 20000
15000
10000
Nooa 5000
Si 2p .I.
.I.
I.
.I.
.I.
..
Binding Energy (eV)
Fig. 8-X-ray photoelectron spectrum of fracture bonding with chemically-cured LOGI.
295.0
BINDING ENERGY (eV)
275.0
....
site
on
dentin after
Fig. 9-Carbon is spectra for (a) dentin, (b) LCGI bond area, and (c) chemically-cured LCGI bond area. The moieties responsible for the various peaks are indicated in (a) and (b) and are shown as dotted lines.
Downloaded from jdr.sagepub.com at Bobst Library, New York University on April 20, 2015 For personal use only. No other uses without permission.
11 ADHESION STUDIES ON GLASS-IONOMER CEMENTS V61.Vol 7171 No.No.11
1839 13
DON
16
A
000000000000000
O
000* 0
0 00000og'
6
N
c F Al Si
0 A p
000a0o
4*
#
C4
v
Z"
z
z
9J
A
K~~~~~~~~~~~~~~~. -
05
DET (JmT
Fig. 10--SIMS depth profile through light-cured glass-ionomer cement and into dentin surface.
ing faied surfaces were examined on both the dentin and the glassionomner sides. This was carried out on half of the samples used in the ahoye bond strength tests. Prior to analysis, all sample surfaces were air-dried and sputter-gold-coated. Xlray phtotoeicctro s pcctroscopy XPS .-XPS analyses were carried out obr detection of chemical changes or the surfaces of the other half of the samples flowing bond strength tests. With an Isomet slow-speed saw, samples were cut to a size of 2 mm x 2 mm. Immediately pnior to analysis by XPS, the samples were evacuated in a vacuum chamber (5 x 10- Pa) so that the water vapor pressure would be reduced to alloy for sample introduction into the UH-V spectrometer analysis chamber. The spectrometer used was a customized SSL SSX-100 x-ray photoelectron spectrometer, which uses a monochromatized Al Ku x-ray gun to excite photoelectrons fromr a spot size assmall as 150 gim in diameter. It is possible for the position of the spot wV ere analysis is to take place to be adjusted with aprecisionr of~t 10 gm. Control specimens ofdentin were always analyzed so that the effect of dehydration of samples in the vacuum chamber of the spectrometer could be monitored. For each sample analyzed, broaascans were accumulated for detection of anl elew ments. During these analyses, a spot size of 600 jim and a pass energy of 148. 0 eV were used. For some ofthe peaks, high-resolution spectra of these particular lines (mainly carbon, oxygen, and nitrogen) were obtained so that chemical effects on photoelectron peak positions could be assessed. For these analyses of the core level spectiad a 150 jim spot size and a pass energy of 50.0 eV were used. The charging effect observed during the photoelectron analysis of non-conducting surfaces was controlled by utilization of an electron flood gun and a metal grid placed close to the surface. Secondary ion mass spcctronmetry (SIMS).In order for the interface between the light-cured glass-ionomer cement and the dentin surface to be analyzed, SIMS depth profiles were camrid out through these layers in a prepared sample. Depth profiles were carried out on light-cured glass-ionomer cement, polished dentin, and interfaces between the two substrates. A Cameca IMS-3f secondary ion mass spectrometer was used. An 0- primary beam provided a stable charge condition on the non-conductive samples. Positive s.econdary jons~were monitored, which provided good ion yields for the elements ofinterest. The elements monitored were H1, 12C, i9F, 2WA, 28Si, 31P, "Ca, and 4Zn. Laser scanning confcal microscopy Confocal microscopy was used for the obtaining of thin optical sections below the surfaces of the specimens. This technique provides a unique opportunity for images of the penetration and thickness of the cement and adhesive in the dental restoration to be obtained. Freshly extracted third molar teeth were used for the restorations. Class V cavities were prepared as in normal clinical practice. All cavities were 2 mm wide and 2 mm high and were cut with a type 012 diamond bur running wet in an ultra-high-speed handpiece. Enamel surfaces were etched
Fig. 11!Scarnnirg electron mvicrograph offracture surface after boarding with light-cured glass-irnomenr ceet, showirng rei tags that have penetrated dentirnal tubules,
with 37% phosphoric acid gel for I5 s and then rinsed with distilled water for 30 s Light-cured glass-ionomer cement (Vitrebond, 3M) and conventional glass-ionomer cement (Glass lonomer Cement Liner, 3M) were both prepared according to the manufacturer's instructions. A thin layer of ionomner was placed over the dentin and light-cured if necessary. Dentin primer (Scotchprep, 3M) was applied, to any exposed dentin surfaces, with a sponge for 60 s and was agitated so that the surface would be kept wet for the time period. The primer was ther air-dried. The adhesive resin (Scotchbond 2, 3M) was then applied to the glass-ionoirer cement, primed dentin, and etched enamel surfaces. The adhesive was 1 ghtcured for 20 s. The cavities were then filled with a composite material (P-SO, 3M) and light-cured. The restored teemh samples were kept tor 24 h in aistiiled water at 37 C With an Isomet slow-speed saw, samples were sectioned longitudinally through the centers of the restorations, and the diamond blade was lubricated with water. Thus, the interface was exposed for examination with the microscope. Five restorations of each type (light-cured vs. chemically-cured glass-ionomer cements) were prepareed, giving a total of~teii block faces. To aid in the visualization of the penetration of the glassionomer cement and the adhesive, fluorescent labels were incorporated into the materials. Double-labeling by fluorescing isothiocyanate (FITC) and Texas Red (Molecular Probes Inc., Eugene, OR) was used. The FITC is maximally excited by the 488-nm (hlue) laser line and emits at wavelengths3 to~vhich the. photonmultipliers are highly sensitive. Texas Red is maximally excited at the 514 nm (green) laser line, thus producing perfect, registration for two-channel confocal imaging. FITC was used for labeling the Scotchbond 2 adhesive resin, and Texas Red was used for labeling the polymer portion of the glass-ionomer cement. This produced the best double-labeled images. Scanning images were obtained in a Bio-Rad MRCC 600 laser scanning confocal microscope (Bio-Rad Inc., Camabridge, MA). The scanned samples were examined for gap formations, penetration of materials into the tooth surfaces, and distribution ofthe two materials relative to each other. Images were recorded with a 35-mm camera with T-MAX 100 black-and-white high contrast film (Kodak).
Results. The bond strength results are presented in Fig. 1. The glassionomer cement samples which were light-cured had the highest
Downloaded from jdr.sagepub.com at Bobst Library, New York University on April 20, 2015 For personal use only. No other uses without permission.
1840
LIN et al.
J Dent Res November 1992
bond strength to dentin, both immediately after curing and following 24 h in ambient conditions (p < 0.001). When the same glassionomer cement was only chemically-cured, the bond strength to dentin was lower than that after light-curing, both immediately and after 24 h (p < 0.001). When the glass-ionomer cement was first allowed to cure chemically for 15 min and was subsequently lightcured, 24-hour bond strengths to dentin began to approach the strength of the light-cured glass ionomer to dentin. An SEM examination of the fracture site suggested a bonding mechanism and the nature of bond failure. For the light-cured ionomer, the bonded area showed the presence of a thin layer ofglass ionomer on the surface of the dentin, as seen in Fig. 2. The lightcured ionomer samples which were chemically-cured resulted in a bonded area with very little residual glass ionomer (Fig. 3). However, a smear layer ofdentin debris was observed on the bonded area ofthe glass-ionomer side, showingthat this system failed cohesively within the smear layer. Scanned images of the sections taken with the confocal microscope showed close adaptation of the light-cured glass-ionomer cement to the dentin surface. Penetration of the polymer into the dentinal tubules is clearly seen in Fig. 4. The chemically-cured glass-ionomer cement exhibited less effective penetration into the dentin surface (Fig. 5). The labeled cement had not moved down along the dentinal tubules to the same extent as had the light-cured material. By the sections being double-labeled, the distribution of both the adhesive and ionomer components of the restoration system could be observed. No adhesive penetrated the ionomer layer and smear layer to enter the tubules, except where ionomer
intensity in the control ionomer when compared with the ionomer close to the interface. Near the interface, the glass ionomer had increased amounts of carbon, hydrogen, fluorine, phosphorus, and zinc. SIMS depth profiling confirmed the ionic exchange process between the light-cured glass-ionomer cement and the dentin surface. There was evidence for the movement of ions from the glassionomer cement 1.5 gim into the dentin surface and the movement of calcium and phosphorus ions from dentin 1.0 pm into the ionomer.
Discussion. When glass ionomers come in contact with a dentin surface, the polymer is believed to bond chemically to the surface with the
production of strong adhesive bonds between the two surfaces. When any type of tangential motion is attempted, shear that occurs in the polymer indicates that the interfacial bond is stronger than the cohesive bond in the polymeric material. The bonds in the cohesively-weaker ofthe two materials will fracture preferentially, with the adhesive bond at the interface being stronger than forces in either of the two materials in contact. When sliding is initiated between the surfaces for the light-cured glass ionomers in contact with the dentin or enamel, shear takes place in the ionomer, and an ionomer film is left on the solid surface. In other words, the interfacial bond formed between the ionomer and the solid substrate is stronger than the cohesive bond in the polymer, and therefore, shear occurs in the glass ionomer. When the material was chemically-cured, the glass-ionomer cement fractured at the smear layer. Failure occurred at the weaker adhesive bonds which were was not present. formed at this ionomer/smear layer interface. The XPS results are presented in Figs. 6, 7, and 8. Those dentin Additional stability of the glass-ionomer surface bond is prosurfaces with smear layer primarily consisted of carbon, oxygen, vided by mechanical interlocking of polymer in dentin. Mechanical nitrogen, calcium, and phosphorus. The dentin side of the failed adhesion would naturally be enhanced by greater contact with the bond area for light-cured glass-ionomer cement to dentin appears to irregularities or pores of the substrates. The viscosity of the attract additional silicon, aluminum, fluoride, and zinc ions, pre- materials is therefore important with respect to this particular sumably from the glass-powder component. Calcium and phospho- mechanism. This type ofmechanical interlocking was evident from rus concentrations are also reduced in this area. The dentin side of SEM observations ofthe bond fracture surfaces and scanned confothe failed bond area for chemically-cured glass-ionomer cement cal microscope images of the interface between the ionomer and which was chemically-cured showed mostly carbon and oxygen and dentin. Fig. 11 is amicrographofthe failed bonded surface ofdentin an even smaller amount of calcium and phosphorus. However, there after fracture. Many tubules are occluded with the light-cured glass was no indication of an increased amount of glass particles, since ionomer, and ionomer resin tags protrude from this surface. The only a small amount of silicon was detected at this surface. restoration section images obtained by the confocal microscope High-resolution spectra for carbon are shown in Fig. 9 and clearly showed penetration ofthe light-cured glass-ionomer cement compared for surfaces of dentin and the failed bond areas of dentin through the smear layer and into the dentinal tubules. for light-cured and chemically-cured glass-ionomer cement. When Rates of adsorption on the calcified tissue surface and crossthe chemical states of carbon from the C(ls) spectra for the dentin linking in bulk regions of the material will be different for the lightsurface, light-cured glass-ionomer bond fracture site, and chemi- and chemically-cured glass-ionomer cements. The ionized groups cally-cured glass-ionomer bond-fracture site were compared, car- compete for sites on the surface and for available cations for crossbon contribution from different functional groups could be detected linking during setting. The hydrolytic stability of the sonically at each surface. Singly-bonded C-OH or C-O-C groups were ob- cross-linked network formed after the material has set, and its served on the surface. Polar groups bonded to carbon were expected, adhesive strength to the hydrophilic dentin surface is the result of due to the nature of the wet dentin surface. Carbonyl groups are a balance ofelectrostatic forces (repulsive and attractive) duringthe present on the surface of dentin and are due to the collagen content course ofthe polyelectrolyte's ionization through an acid/base reacin the dentin smear layer. A higher amount of carboxyl groups was tion. For example, rapid ionization once the cation source and the detected on the chemically-cured-ionomer failed-bond site com- polymer are brought into contact might lead to the adsorption of a pared with that on the light-cured ionomer. This indicates a low thin, strongly bound layer of polymer material on dentin. This layer amount of ionic bonds present where the bond failed for the light- does not necessarily allow for a molecular continuity between the cured material. Chemical analysis of interfaces showed that differ- tissue interface and the bulkofthe sample (Belton and Stupp, 1988). ent bondingmechanisms may have been responsible forthe bonding Light-curing causes an initial increase in the ionization rate due to of the light-cured and chemically-cured systems. the production of acid (via the photoinitiator), and this results in a SIMS depth profiles are best compared by ionomer and dentin very strong adsorbed layer on dentin. At the same time, internal interfacial profiles being overlaid with the control profiles of glass- cross-linking reactions result from free-radical-polymerization proionomer cement and dentin, as in Fig. 10. On the control dentin cesses. Thus, the material itselfwill have secondary bonding in the surface, calcium, phosphorus, and hydrogen were of intensities bulk among molecular or between side groups which can then similar to those in dentin near the interface. Carbon, aluminum, prevent the cations from approaching the carboxyl groups for crosssilicon, and fluorine were sharply higher at the surface ofthe dentin linking. near the interface. Zinc intensities were moderately higher at the For the chemically-cured glass-ionomer cement, the reduced interface. Aluminum, silicon, fluorine, and zinc were of similar number of carboxylate groups in the polymer due to the presence of Downloaded from jdr.sagepub.com at Bobst Library, New York University on April 20, 2015 For personal use only. No other uses without permission.
Vol. 71 No. 11
ADHESION STUDIES ON GlASS-IONOMER CEMENTS
side groups can decrease the amount of adsorption on the surface of the dentin. From the SEM and correspondingXPS results, it is clear that light-curing and the chemical structures of the polymer used will significantly affect the adhesion characteristics of the glassionomer cement. Further investigation is needed for better understanding ofthe role ofionization rate on adhesion properties and the driving forces for both chemical bonding and mechanical retention to the dentin surfaces. The combined results from the above experiments have provided a more comprehensive view ofthe interfaces between glass-ionomer cements and dentin substrates. Based on these preliminary studies, further investigations on the nature ofchemical bonding at the interface and the role ofmechanical interlocking would be useful for the purpose of optimizing adhesion to the components of tooth structure in an aqueous environment.
Acknowledgments. We gratefully thank Dr. M. Hubert, Ontario Laser and Lightwave
1841
Centre, University of Toronto, for his skilled technical assistance and use of the scanning laser confocal microscope, and Dr. K Nielsen, 3M Canada Inc., for his assistance with the fluorescence labeling. REFERENCES Belton D, Stupp SI (1988). Adsorption of ionizable polymers on ionic surfaces: poly (acrylic acid). Macromolecules 16:1143-1150. Smith DC (1989). In-vitro performance of glass ionomer cements. In: Proceedings ofthe conference oncorrelationbetweenin-vitro and in-vivo performance of dental materials. Trans Acad Dent Mater (Dublin, Ireland). Watson TF (1989). A confocal optical microscope study ofthe morphology of the tooth/restoration interface using Scotchbond 2 dentin adhesive. J Dent Res 68:1124-1131. Wilson AD, McLean .1W (1988). Glass ionomer cements. Chicago (IL): Quintessence Publishing Co. Inc., 13-39.
Downloaded from jdr.sagepub.com at Bobst Library, New York University on April 20, 2015 For personal use only. No other uses without permission.
