J /O A

A R T IC L E S

Etching or roughening the surface of glass ionomer cement before use of composite resins
and bond agents produces bond strengths comparable to the bond strength between glass
ionomers and dentin. Bond failure at such surfaces occurs within the glass ionomer.
Adequate washing with water after acid etching the glass ionomer is essential to obtain
optim al bond strength. Apparently, some combinations of ionomer cements and resins are
more effective than are others in providing a good bond in the “sandwich technique. ”

Tensile bond strength between

glass ionomer cements and composite resins

Ko H inoura, DDS, DDSc

B. Keith Moore, P h D
R alp h W. P h illip s, MS, DSc

he adhesion of glass ionom er etched. T he area then is restored with a depends on achieving a reasonable bond

T cements to enamel and dentin1-4

has resulted in the use of glass ion­
omer cements to restore cervical abrasion
composite resin. The procedure makes
optim al use of the adhesive properties and
biocom patibility of the glass ionomer
strength between the etched glass ionomer
cement and the composite resin and its
bonding agent. Although enamel acid etch­
or erosion lesions5' 8 and to treat dentin cement and the desirable surface and es­ ing improves the bonding of resin to that
hypersensitivity.9 Bonding to enamel and thetic appearance of the composite resin. surface, the bond strength of composite
dentin requires little tooth preparation. As (This concept was first demonstrated in a resin and bond agents to etched glass ion­
with silicate cement, fluoride release is course presented in 1976 by Drs. John W. omer has not been reported. T his study
available from glass ionomer cements for a McLean and R alph W. Phillips.) evaluates the bond strength between sev­
prolonged period.2'10’11 Therefore, a caries- T he success of the sandwich technique eral glass ionomer cements and composite
preventive effect can be expected. In con­
trast to pulp reaction to silicate cement, the
p ulp reaction to glass ionomer cements Table 1 ■ Glass ionomer cements evaluated.*
appears m ild.12-14 P ow der-to-liquid
Because teeth restored with glass iono­ G lass io n o m er cem ent M anufacturer m ix in g ratio B atch no.
mer cements usually are less pleasing in G C lin in g cem ent G C , Jap a n 1.2 g:1.0 g 060351/040351
appearance and are less resistant to abra­ F u ji Io n o m er T ype I G C , Jap a n 1.35 g:1.0 g 280111/210111
sion than are composite resins, the “sand­ F u ji Ionom er T ype II GC, Japan 2.3 g:1 .0 g 040451/290351
M iracle M ix G C , Jap a n 6.03 g:1.0 g 040351/250351/041285
wich technique” was developed. Although Ketac-bond ESPE-Prem ier 3.4 g:1.0 g M092MD040285
various modifications of the sandwich Ketac-silver ESPE-Prem ier 1.35 g:0.25 m L M039MD020885
technique exist, the technique can be sum ­
•P ublication of names of products does not imply endorsement by the American Dental Association.
marized as follows. Standard Class I, II, III,
or V cavities are cut. After all caries is
removed, the cavity is filled with glass ion­
omer cement. Next, the enamel walls are Table 2 ■ Composite resins and bonding agents used.
cleaned of cement and the enamel and C om posite resin B onding agent M anufacturer B atch no.
cement are etched with phosphoric acid. A M icrorest AP G C Bond GC, J a p a n 101241
bonding agent is applied, followed by a V isio-dispers V isio-bond ESPE-Prem ier M023MD012385
composite resin. In an eroded area, only a S ilux Scotchbond 3M Co 031485

thin layer of the cement is applied and

JADA, Vol. 114, February 1987 ■ 167

ARTICLES

Fig 1 ■ Cross-section of the apparatus used to

measure the tensile bond strength. Key: A = a
threaded cap with wires attached for the insertion
in grip of the testing machine; B = a two-component
mold (Delrin) with a 30° chamfer; C = resin or glass Fig 3 ■ Scanning electron microscope slides of Ketac-silver surfaces after treatments (orig mag X 200). Top
ionomer cement. left, smooth surface. Top right, etched surface with 30-second wash. Bottom left, etched surface with
5-second wash. Bottom right, ground surface.

00104

Fig 2 ■ Scanning electron microscope slides of GC lining cement surfaces after treatments (orig mag X 200). Left, smooth surface. Middle, etched surface. Right,
ground surface.

168 ■ JADA, Vol. 114, February 1987

 resins with their bond agents and compares
various surface treatments of the cement.

Methods and materials

T h e bond strength was determ ined by subjecting

paired cylinders of the m aterials to a tensile-type
stress. A tw o -p art m old (D elrin), 12 m m in
len g th a n d 4 m m in diam eter (Fig 1), was used to
form a n d h o ld the cem ent a n d resin. A 30°
cham fer was m achined at one end of each h alf of
the m old to re ta in the m aterials d u rin g loading.
Six glass ionom er cem ents were evaluated (T able
1). T h e G C lin in g cem ent, F u ji Ionom er T y p e I,
F uji Ionom er T y p e II, M iracle M ix, a n d Ketac-
b ond were m ixed o n the p ap er pad as recom ­
m ended by the respective m anufacturer. Ketac-
silver w as m ixed in a n am alg am ato r (Vari-M ix
II M) for 10 seconds a t a n M-2 setting, as recom ­
m ended by the m anufacturer.
T h e glass io n o m er cem ent was inserted in to
Fig 4 ■ Scanning electron microscope slides of Miracle Mix surfaces after treatments (orig mag X 200).
the 6- X 4-m m m o ld a n d placed a g ain st a flat
Left, smooth surface. Right, etched surface.
glass p late o n the end opposite the cham fer.
After the glass io n o m er cem ent h a d set for 7
m inutes at room tem perature, the flat surface of

m Smooth surface H Smooth surface

1 1 Etched surface 0 Etched surface
Ground surface

Microrest AP (GC Bond) Visio-dispers (Visio-bond)

m Smooth surface
□ Etched surface
□ Ground surface
60
Fig 5 ■ Bond strength between Microrest AP/GC
Bond (top left), Visio-dispers/Visio-bond (top
e right), and Silux/Scotchbond (bottom left) and var-
^60 40 H
lonomer cements.

20 -

LC FI F II MM KB KS
Silux (Scotchbond)

H in o u ra -M o o re -P h illip s : BOND S T R E N G T H B ETW EEN GLASS IO N O M E R C E M E N T S AND RESINS ■ 169

ARTICLES

the glass ionom er was treated as follows: surface specim ens, all failures appeared to occur
—G ro u p S: n o treatm ent, glass ionom er cohesively in the cem ent (Fig 6).
cem ent sets ag ain st glass (sm ooth surface). T h e results were studied w ith analyses of var­
—G ro u p E: glass ionom er cem ent sets against iance follow ed by T u k ey ’s tests. In T ables 3-7,
glass, is etched w ith 37% p h o sp h o ric acid for 60 values connected by vertical lines are n o t sig n ifi­
seconds, a n d w ashed w ith tap water for 30 cantly different (P < .01).
seconds (etched surface). A 60-second etch was T a b le 3 show s the effect of the surface treat­
used to assure a m axim um etch pattern. H o w ­ m ents of the glass ionom er cem ent surfaces on
ever, in the clinical use of this technique, 30 the bond strength for all cem ents and resins. T h e
seconds is adequate, p articularly w hen the layer e tching o r m echanical ro u g h e n in g of the surface
of cem ent is thin. of the cem ent had a significant effect on the bond
—G ro u p G: glass ionom er cem ent sets against strength. T h e adhesion between glass ionom er
glass a n d is ro ughened w ith 400-grit silicon car­ cem ent a n d com posite resin was enhanced either
bide p ap er (ground surface). by acid e tching or ro u g h en in g of the cem ent
O ne h alf of the m old was a ligned w ith the surface. E tching resulted in a modest increase in
o th er half, w hich was filled w ith cem ent and the bond strength as com pared w ith the ro u g h ­
held in po sitio n w ith plastic electrical tape. ened surface.
T h ree m icrofilled com posite resins were used T h e effect of the com posite resin a n d its b o n d ­
(T able 2). M icrorest AP, a chem ically cured in g agent on the bond strength is show n in
resin, w as used w ith its chem ically cured enam el T ab le 4, w hich includes all the cem ents and
b o n d in g agent. T h e bon d in g agent was m ixed surface treatm ents. S ilux/S cotchbond had the
Fig 6 ■ Scanning electron microscope slide/frac­
an d placed over the cem ent, a n d the resin was highest bond strength, V isio-dispers/V isio-bond
ture showing failure in the cement (cohesive fail­
m ixed a n d b u lk placed to fill the m old. T h e had the next highest bond strength, a n d M icro­
ure) surface between Ketac-silver and Siiux (orig
open end was covered w ith a m atrix strip, and rest A P /G C B ond had the lowest bond strength.
mag X 35).
the specim en was allow ed to set for 1 hour. T h e effect of the glass ionom er cem ents on the
V isio-dispers a n d Silux and th eir respective bond strength is show n in T ab le 5. Ketac-silver
b o n d in g agents are lig h t cured. In each case, the an d M iracle Mix show ed the highest bond
b o n d in g agent was placed and cured for 20 m ens were bonded to Visio-dispers a n d tested as strength irrespective of the resin or surface
seconds u sin g a Kerr C om m and light. N ext, the described. treatm ent used, Fuji I and F uji II show ed m od­
com posite resin was condensed in to the m old in T o determ ine the cohesive strength of the erate bond strength, and K etac-bond a n d G C
three increm ents, each of w hich was cured for 30 individual glass ionom er m aterials, the tensile lin in g cem ent showed the lowest b ond strength.
seconds. T h e finished specim ens set for 1 hour, strength of 24-hour-old specim ens was deter­ T ab le 6 shows the re lationship betw een the
at w hich tim e all assem bled specim ens were m ined by the app licatio n of a tensile load to various com posite re sin s/b o n d in g agents and
transferred to distilled water and stored at 37 C. cylinders of the m aterials. T h e m old a n d tech­ the surface treatm ents of the cem ent. Significant
After 24 hours, the electrical tapes were re­ n iq u e used for these experim ents have been des­ differences existed between etched, g ro u n d su r­
m oved a n d the m olds were screwed in to the cribed.15 T h e 24-hour tensile strengths were face, a n d sm ooth surfaces w ith the exception of
threaded caps to attach them to the testing determ ined (Instron T esting M achine) at a cross­ the S ilux/S cotchbond system, w hich show ed no
m achine. T en sile bond strengths of seven spec­ head speed of 0.5 m m /m in . significant difference between the sm ooth and
im ens for each cem ent-resin com bination were gro u n d surfaces.
determ ined (Instron T estin g M achine) w ith a T hese data also m ay be analyzed to show the
Results
crosshead speed of 0.5 m m /m in . re la tio n sh ip between surface treatm ents and
T h e effect of the acid e tching on the surface of com posite resins in cluding all glass ionom er
the glass ionom er cem ent also was studied on S canning electron m icroscope pictures for the cem ents (T able 7). O n a sm ooth cem ent surface,
positive replicas usin g a scanning electron m i­ various glass ionom er surface treatm ents are the b ond of S ilux/S cotchbond was significantly
croscope. Im pressions of the surfaces were taken show n in Figures 2-4. Figure 5 shows the tensile stronger th an w as the bond of M icrorest A P /G C
w ith Iight-body polyvinyl siloxane im pression bond strength between three com posite resins b o n d a n d Visio-dispers/Visio-bond. F o ra ll types
m aterial (M irror 3), and a positive replica poured and six glass ionom er cem ents w ith various su r­ of surface treatm ent, bond strengths of S ilu x /
w ith epoxy (Stycast 1266). face treatm ents. Bond strengths varied from 61.1 Scotchbond and V isio-dispers/V isio-bond were
T h e effect of varying the w ashing tim e after k g /c m 2 for Ketac-silver, etched, w ith S ilu x / significantly stronger th an was the bond strength
e tching the glass ionom er cem ent was also stu d ­ Scotchbond to 1.4 k g /cm 2 for unetched G C lin ­ of M icrorest A P /G C Bond.
ied. Ketac-silver was allow ed to set against glass in g cem ent bond to V isio-dispers/V isio-bond. T ab le 8 shows the effect of the w ashing tim e
for 7 m in u te s a n d th en was etched w ith 37% W ith the sm ooth surface glass ionom er spec­ after e tching of the cem ent. T hese data were
p h o sp h o ric acid for 60 seconds a n d w ashed w ith im ens, all failures occurred adhesively at the analyzed by S tudent’s i-test. A significant differ­
tap w ater for either 5 or 30 seconds. These speci- c em ent/bonding agent interface. W ith the etched ence was found between the 30-second a n d 5-

Table 5 ■ Effect of the glass ionomer

Table 3 ■ Effect of the surface treat­ Table 4 ■ Effect of the composite resin cements on the bond strength.*
ments of the glass ionomer cement sur­ and its bonding agent on the bond
G lass ionom er cem ent M ean (k g /c m 2)
faces on the bond strength for all cements strength.
Ketac-silver 41.7 “ I:
and resins. C om posite re sin / M iracle Mix J
41.4 ;
Surface treatm en t M ean (k g /c m 2) b o n d in g agent M ean (k g /cm 2) F u ji Ionom er T y p e II 3 4 .6 “]
Etched surface 37.7 S ilu x /S co tch b o n d 40.6 F u ji Ionom er T ype I 30.7 J .
G ro u n d surface 32.0 V isio-dispers/V isio-bond 32.7 Ketac-bond 17.5“]
Sm ooth surface 21.6 M icrorest A P /G C B ond 17.9 G C lin in g cem ent 16.9 J
•Bond strengths connected by brackets on the right were
n ot statistically different (P < .05).

170 ■ JADA, Vol. 114, February 1987

 ARTICLES

Microrest A P/G C Bond. For the etched or

the ground surface, however, Silux/Scotch­
Table 6 ■ Effect of composite resins and surface treatments.* bond and Visio-dispers/Visio-bond were
C om posite re sin /b o n d in g agent an d surface treatm ent M ean (k g /cm 2) both significantly stronger than was Micro­
M icrorest A P /G C B ond rest A P/G C Bond. This suggests that dif­
Etched surface 24.1 ferent types of chemical bonding may be
G ro u n d surface 19.9 found between glass ionomer cements and
S m ooth surface 9.8
the bonding agents for Silux/Scotchbond
V isio-dispers/V isio-bond
Etched surface 42.7
and Visio-dispers/Visio-bond systems, in
G ro u n d surface 36.6 addition to any subtle differences in viscos­
S m ooth surface 19.0 ity of the bond agents. These results may
S ilux/S cotch b o n d also be related to the tensile properties of
Etched surface 46.2
the three composite resins and their bond­
G ro u n d surface 39 T \
Sm ooth surface 3 6 .JJ ing agents.
•Bond strengths connected by brackets on the right were not statistically different (P < .05).
There are even significant differences in
bond strengths between the glass ionomer
cements regardless of the resin used or the
surface treatment. Ketac-silver and Miracle
Mix showed the highest bond strength,
Table 7 ■ Effect of surface treatments and composite resins.*
Fuji Ionomer I and Fuji Ionomer II were
Surface treatm ent an d com posite resin Mean (k g /c m 2) classified into an intermediate group, and
Sm ooth surface Ketac-bond and GC lining cement were
Silux 36.1 classified into the lowest group. Ketac-
V isio-dispers 19.0
silver and Miracle Mix surfaces set against
M icrorest A P 9.8
Etched surface glass show many small bubbles (Fig 3, 4).
Silux 46.21 These small holes on the surface could con­
V isio-dispers 4 2 .7 1 tribute to the m echanical interlocking
M icrorest AP 24.1 between the cement and the composite
G ro u n d surface
S ilux 39.6"l
resin.
V isio-dispers 3 6 .6 1 It is apparent that the washing time after
M icrorest A P 19.9 etching is im portant to the bond strength.
•Bond strengths connected by brackets on the right were not statistically different (P < .05). If washing time is too short, the debris
rem aining on the surface of the etched
cement results in an adhesive failure.
second w ashing periods. enter into the irregularities on the rough T he shear bond strength of glass iono­
T ab le 9 show s the tensile strengths of the six surface. Fukushim a16 reported that acid mer cement to dentin has been reported as
glass io n o m er cem ents. W hen the results were etching and applying low-viscosity bond­ approximately 2.5 mPa in vitro.14 Con­
analyzed by N ew m an-K euls test w ith critical ing agents to dentin appears to be im por­ versely, Powis and others3 found that the
values at .05, they separated in to two groups.
tant in enhancing the resin’s adhesion to
T h e h ig h er tensile strength g ro u p includes Fuji
dentin. The three bonding agents used in
Ionom er T y p e II, Ketac-silver, M iracle M ix, and
F uji Ionom er T ype I; b u t Ketac-bond a n d GC
this study had low viscosity. Apparently, Table 8 ■ Effect of the washing time
lin in g cem ent were included in the low er tensile the same principle applies in this tech­ after etching of the cement (N - 7).*
strength g ro u p w ith the overlap g ro u p being nique; acid etching of the cement followed
W ashing tim e__________ K g/cm 2_________(SD)
Fuji Ionom er T ype I a n d Ketac-bond. by application of a low-viscosity bonding
E tched and
agent is an im portant means of ensuring
w ashed for 30 sec 50.4 (12.49)
Discussion bonding to the composite resin. E tched and
Sneed and Looper17 reported that the w ashed for 5 sec________30.7________ (12.71)
bond between an etched glass ionomer •T h e different washing times were compared using the
Etching the surface of glass ionomer cement cement and a composite resin is stronger Student’s ¿-test.
They were significantly different at t = 2.92, d f= 12.
greatly increased the bond strength to all than is the cohesive strength of the glass
the composite resins/bonding agents stud­ ionomer cement. Their findings are sim­
ied. The matrix of the hardened glass ion­ ilar to the data presented in this report. An
omer cement dissolves in acid, resulting in example of a cohesive failure is shown in Table 9 ■ Tensile strength of glass ion­
a rough and porous surface. The bonding Figure 6. omer cement (N = 3).*
agent then can penetrate into the surface These results indicate that Silux/Scotch- G lass io n o m er cem ent K g /c m 2 (SD)______
irregularities and harden. Specifically, the bond showed the highest bond strength F uji Ionom er T ype II 48.0 (4.53H
unfilled resin infiltrates the etched porosi­ irrespective of glass ionomer cements or K etac-silver 47.5 (3.22)
ties and during polymerization creates re­ surface treatments. Next was Visio-dispers/ M iracle Mix 45.4 (7.08)
tentive tags at the interface. F uji Ionom er T ype I 35.3 (1.88)
Visio-bond, and Microrest A P/GC Bond
K etac-bond 29.0 (9.82)
Bond strength is reliant on many factors, had the lowest strength. When comparing G C L in in g cem ent 18.8 (5.52)
one of which is contact angle. On a rough the bond strength on the smooth surface,
•Strengths connected by brackets on the right were not
surface the contact angle should be low so Silux/Scotchbond was significantly strong­ statistically different (P 5~ .05).
that the bonding agents can spread and er than was Visio-dispers/Visio-bond and

H in o u ra -M o o re -P h illip s : BOND S T R E N G T H BETW EEN GLASS IO N O M E R CE M E N T S AN D RESINS ■ 171

ARTICLES

tensile bond strength of glass ionomer reported between glass ionomer cements study. Oper Dent 7(l):2-6,1982.
5. McLean, J.W., and Wilson, A.D. The clinical
cement to dentin is between 1.8 and 7.3 and dentin. The bond strength values mea­
development of the glass-ionomer cements. Formula­
mPa. In this study, the bond strengths sured in this study support the use of glass tions and properties. Aust Dent J 22( 1);31 -36, 120-127,
between the etched glass ionomer cements ionomer materials as an intermediate layer 190-195, 1977.
and composite resins are between 0.9 and between tooth structure and an overlying 6. Mount, G. J., and Makinson, O.F. Clinical charac­
6.1 mPa. These bond strengths seem to be composite resin in the sandwich technique. teristics of a glass-ionomer cement. BrDent J 145(3):67-
71, 1978.
adequate when compared with the data ---------------------- JM O A ---------------------- 7. Lawrence, L.G. Cervical glass ionomer restora­
cited earlier on the glass ionomer-dentin tions: a clinical study. Can Dent Assoc J 45(2):58-59,
bond. 1979.
Some combinations of ionomer cements Information about the manufacturers of the products 8. Phillips, R.W. The restoration of eroded cervical
mentioned in this article may be available from the areas. CDS Rev 73(4):31-34,1980.
and resins appear to be more effective than author. Neither the authors nor the American Dental 9. Onose, H. The treatment of dentin hypersensitiv­
others in providing a good bond. More Association has any commercial interests in the pro­ ity by the glass ionomer cement mix application. J Int
research is needed to further determine the ducts mentioned. Coll Dent 14(l):29-35, 1984.
factors involved in achieving maximum 10. Swartz, M.L., and others. Fluoride distribution
The authors thank Hazel E. Clark for assistance with in teeth using a silicate model. J Dent Res 59(10):1596-
bond strengths between etched ionomer
the analysis of the data. 1603, 1980.
cements and the respective bonding agent- 11. Swartz, M.L.; Phillips, R.W.; and Clark, H.E.
resin system. Dr. Hinoura is instructor, department of operative Long-term F release from glass ionomer cements. J
dentistry, Nihon University, Tokyo, and visiting pro­ Dent Res 6S(2):158-160,1984.
Summary fessor; Dr. Moore is professor; and Dr. Phillips is asso­ 12. Tobias, R.S., and others. Pulpal response to a
ciate dean for research and research professor, depart­ glass ionomer cement. BrDentJ 144(ll):345-350,1978.
ment of dental materials, School of Dentistry, Indiana 13. Kawahara, H.; Imanishi, Y.; and Oshima, H.
Etching or surface roughening of a glass University, 1121 W Michigan St, Indianapolis, 46202. Biological evaluation of glass ionomer cement. J Dent
ionomer cementbefore the application of a Address requests for reprints to Dr. Phillips. Res 58(3):1080-1086, 1979.
composite resin bonding agent leads to a 14. Pameijer, C.H.; Segal, E.; and Richardson, J.
significant improvement in the resulting 1. Sato, S.; Sato, K.; and Hosoda, H. Adhesion of new Pulpal response to a glass-ionomer cement in primates.
dental cements. J Japan Res Dent Mat Appliances J Prosthet Dent 46(1 ):36-40, 1981.
bond. Bond failure at the surfaces occurs 35(l):68-79, 1978. 15. Phillips, R.W., and others. Zinc oxide and
cohesively within the glass ionomer. Ade­ 2. Maldonado, A.; Swartz, M.L.; and Phillips, R.W. eugenol cements for permanent cementation. J Prosthet
quate washing with water after acid etch­ An in vitro study of certain properties of a glass iono­ Dent 19(2):144-150, 1968.
ing the glass ionomer is essential to obtain mer cement. JADA 96(5):785-791, 1978. 16. Fukushima, M. Tags of various restorative resins
3. Powis, D.R., and others. Improved adhesion of a penetrated into dentin. Japanese J Conserv Dent
optimal bond strengths. Properly treated
glass ionomer cement to dentin and enamel. J Dent Res 22(3):103-120, 1979.
glass ionomer surfaces result in bond 61(12):1416-1422, 1982. 17. Sneed, W.D., and Looper, S.W. Shear bond
strengths to composite resins and their 4. Coury, T.L., and others. Adhesiveness of glass- strength of a composite resin to an etched glass iono­
bonding agents comparable to the strengths ionomer cement to enamel and dentin; a laboratory mer. Dent Mater J 1(4):127-128, 1985.

172 ■ JADA, Vol. 114, February 1987