Destructive Stresses in Adhesive Luting Cements C.L. DAVIDSON, L. VAN ZEGHBROECK', and A.J. FEILZER Department of Dental Materials Science, ACTA, Louwesweg 1, 1066 EA Amsterdam, The Netherlands, and Catholic University of Louvain, Belgium
In this study, curing shrinkage stress development was monitored in a glass-ionomer and a BisGMA composite luting cement adhesively placed at film thicknesses ranging from 30 to 200 ALm. The nature and magnitude of the stress development depended greatly on the formulation and film thicknesses of the lute. The thicker the layer, the faster the stress development in the glass ionomer and the slower in the composite. The contraction stress had a detrimental effect on the cohesive strength of the glass ionomer and on the adhesive strength of the composite cement.
J Dent Res 70(5):880-882, May, 1991
Introduction. When the contraction of adhesive materials is obstructed, stresses are induced which can be detrimental for the cohesion of the structure (Bowen et al., 1983; Davidson et al., 1984). In particular, in cases with an unfavorable configuration-that is, where the ratio of bonded surface to free surface of the shrinking material is higher than five-all shrinkage is directed uniaxially, leading to a contraction that is almost three times higher than the expected linear shrinkage (Feilzer et al., 1989). Cement films will readily surpass such values. Moreover, tensile contraction stresses can exceed the adhesive or cohesive strength of the materials involved (Feilzer et al., 1987). Because all adhesive dental luting cements shrink considerably during setting, it is of interest to investigate the extent to which these cements fail when used as luting agents in rigid constructions. The purpose of the present study was to measure stress development in cement layers of various thicknesses and to demonstrate evidence of destruction of the lutes.
Materials and methods. One glass-ionomer cement [Fuji I-New, GC Int. Co., Japan, batch 130771(p) and 100771(1)] in a powder/liquid ratio of 1.8/1 and one BisGMA-based composite cement (Panavia Ex, Kuraray, Japan, batch no. 43121) in a powder/liquid ratio of 3.2/1 were tested. Stress development was determined at 24 PC in a setup as described by Feilzer et al. (1987), in which two opposing, identical, parallel stainless-steel disks (diameter = 10 mm) were connected to the load-cell and cross-head of a tensilometer (Instron Co., Bucks, UK). The axial sample contraction was continuously neutralized by a feedback displacement of the cross-head in order to maintain the original disk-to-disk distance with an accuracy of 0.2 pum. The steel disks were sandblasted with 250-ptm particles (Bego, Germany) and tin-plated (Kura-Ace tin plating, Kuraray Co., Ltd., Japan). This surface treatment ensured a bond strength that exceeded the cohesive strength of the materials under investigation. The cements were handled according to the manufacturer's instructions and were Received for publication June 13, 1990 Accepted for publication January 7, 1991 880
'Department of Prosthodontics,
inserted between the disks. Optimal adaptation was ensured by squeezing surplus material between the two disks until the desired disk-to-disk distance was obtained. Excess material was removed at the circumference of the disks. The glassionomer margin was coated with vaseline to prevent dehydration, and the composite margin was covered with Oxyguard® (as advised by the manufacturer) to promote complete setting. Two minutes from the start of mixing, continuous shrinkagestress recordings were made at pre-set disk-to-disk distances of 30, 50, 100, and 200 pum. At each film thickness, the average of five recordings was taken. So that the influence of the rigid set-up on the tensile strength of the cements could be determined, similar samples were tested at a cross-head speed of 0.5 mm/min at intervals of one min from four to 30 min after setting. Such experiments were done in a rigid set-up, as well as one in which the cements were allowed to shrink freely into the axial direction. Each experiment was repeated four times, and the results were averaged. The glass ionomer was tested at a film thickness of 30 pum, and the composite at 50 WLm, the minimum film thickness of the material (Van Zeghbroeck, 1989). In many cases, the samples broke spontaneously within the duration of the experiment. Non-linear regression analysis was carried out, and the significance of differences was determined at the 5% level by Student's t test. The measurement of the bond strength to dentin was carried out in the set-up which allowed free shrinkage into the axial direction. A series of dentin samples was cemented to a tincoated stainless-steel disk with the glass-ionomer luting cement in thicknesses of 30 ,um and with the composite in thicknesses of 50 ,um. The strength of the material was measured, at oneminute intervals, from four to 30 min after the start of the mixing.
All the fractured surfaces were examined under a stereomicroscope (X50) (Wild Heerbrugg, Switzerland) to check the nature of the fracture.
Results. Fig. 1 shows representative graphs of the shrinkage stress vs. time for each of the different glass-ionomer cement thicknesses. The nature of the stress development differed with the film thickness. The thicker the cement layer, the earlier the reaction started, and the faster and more continuous was the stress development. In thin layers (30 pum), the stress development showed a pronounced stick-slip contraction pattern. A representative graph of the stress development during setting of the composite is shown in Fig. 2. Unlike the stress development in the glass-ionomer cement, the reaction occurred in the composite cement faster in thinner layers, and the stress pattern was completely different. The setting reaction with this anaerobic setting material started immediately after application of the oxygen barrier and increased continuously until fracture. Non-linear regression analysis plots for the tensile strength as a function of time, for the glass ionomer in the free-shrinking and the rigid set-up, are given in Fig. 3. The correlation coefficients for both cases were, respectively, r2 =
Downloaded from jdr.sagepub.com at UNIV OF MICHIGAN on June 17, 2015 For personal use only. No other uses without permission.
STRESSES IN ADHESIVE LUTING CEMENTS
Vol. 70 No. 5
0.987 and r2 = 0.935 at p0.05). The
early tensile bond strengths of the glass-ionomer and the composite cement to dentin are shown in Figs. 3 and 4, respectively. The glass-ionomer cement always broke cohesively. The composite established its bond strength toward the dentin surface late, in comparison with the hardening stress development. The cement broke in a mixed cohesive/adhesive mode.
881
strength of the cement must develop more rapidly and be at all times superior to the tensile stresses that accompany polymerization. Within five min, the glass-ionomer luting cement had begun to establish a bond with dentin and the alloy. Shortly after gelation, when stress development starts (Bausch et al., 1982), the cement had reached its maximal bond strength to dentin (compare Figs. 1 and 3). Although the bond strength never exceeded 5 MPa, cohesive failure was observed in all cases. In the thinnest layers (30 ALm), the stress development grew gradually in time until a certain point, at which a sudden increase in stress development was registered, followed by the formation of a constant stress level during an interval period of time (see plateau parts in the curves). Although the relative error caused by the resolution of the measuring system increases with decreasing cement layer thickness, this does not explain the observed phenomena, because the resolution of the tensilometer in combination with the displacement measuring system was better than 0.04 jam. In addition, this phenomenon was not seen in the composite cement tested. An alternative explanation is that the glass ionomer forms microcracks due
Discussion.
20r FUJI-NEW 301um
In order to have a clinically useful adhesive bonding between the cement and the preparation and restoration, the bond
free I .m riaid
15
shrinkagege set-up set-un
L
-100 z
early bond streplength to dentin
ul
a:
5
A-- - --
-J
z
2-
FOE
U)
5
I
10
15
20
25
30 time (min)
LU
_
0)
-'._.- '
10
5
15
Fig.
20
25 30 time (min) . Fia 11-1Thp11rb< hrla-rci d IUlIUilUII U1f tLiIIl, fiuV 1-ilUW da; r'g6 iIaIdgr- btlase III IUJI. four different film thicknesses. The ends of the curves represent the points of cohesive failure.
ni..
20
3-The early tensile strength of Fuji I-New
as a
function of time
at a film thickness of 30 prm for a free-shrinkage and a rigid set-up. The shaded area indicates a 95% confidence interval. The development of the bond strength of the cement with dentin is indicated by the dotted line.
The bars
represent the S.D.
T
PANAVIA-EX
|,,,,X,~~ E
15
o1 0
10
E 0-
I I
C12 I I
'
a.
5 LU
5
I--
10
15
20
25
30
0)
5
10
15
20
25
30
time (min)
Fig. 2-The shrinkage stress in Panavia Ex as a function of setting time for three different film thicknesses. The dotted parts of the curves represent
Fig. 4-The early tensile strength of Panavia Ex as a function of time at a film thickness of 50 pum for a free-shrinkage and a rigid set-up. The shaded area indicates a 95% confidence interval. The development of the bond strength of the cement with dentin is indicated by the dotted line.
the range of cohesive failure.
The bars represent the S.D.
time (min)
Downloaded from jdr.sagepub.com at UNIV OF MICHIGAN on June 17, 2015 For personal use only. No other uses without permission.
882
DAVIDSON et al.
J Dent Res May 1991
E
earl -,V
crack formation
stag
,,,
=-rfl
curing contraction
Fig. 5-Schematic representation of crack formation and blunting of the crack tips due to prolonged tensile stress in a not-fully-set glass-ionomer cement. Cracks lead to cohesive failure in more brittle cement.
to the stress development in the incompletely set cement. Microcracks enlarge the free surface and relieve stress (Fig. 5). Due to the continuing setting process, the cement shrinks further and the stress build-up proceeds. However, when the cement ages, it becomes more brittle, so that newly-formed cracks cannot be blunted and stress relief is excluded, resulting in cohesive failure of the cement. The occurrence of plateaus in the stress-time curve had an inverse relation with the cement thickness. In the 200-pLm cement layers, plateaus were not
observed. This can be explained by a more homogeneous distribution of the curing shrinkage stresses in relatively thicker cement layers and the absence of the highly stressed regions that would normally induce microcracking. The abovementioned hypothesis of microcrack formation is supported by the fact that the strength of the cement is decreased by 25-30% in a rigid set-up. Moreover, the existence of microcracks in thin layers of glass ionomer, when cured in a rigid set-up, has been demonstrated microscopically (Van Zeghbroeck et al., 1989). This leads to the conclusion that the glass-ionomer cement is damaged by its own hardening stresses in a rigid set-up. The composite was able to establish a bond strength of almost 5 MPa to the dentin surface, but in contrast to the glassionomer cement, the bond with dentin was not yet formed when the hardening stress developed (Fig. 4). A cement with these hardening and setting characteristics is not likely to establish a clinical bond to the dentin cavity walls. The stress development showed an inverse relation with the film thickness of the composite (Fig. 2). This can be explained by the decreasing flow capacity of the cement with decreasing layer thicknesses. Due to a lower flow capacity in a thin layer, the shrinkage stress develops faster and thus surpasses sooner, and at a lower value, the cohesive strength of the cement (Feilzer et al., 1989). Once the oxygen barrier was placed on the composite, the stress developed continuously until fracture. Because the composite fractured cohesively in this set-up, the tensile strength of the cement could not be measured as a function of time in a completely rigid set-up with the two prepared steel disks. The stress development was allowed to reach only 85% of the total strength of the cement before the stressing was brought to a stop. However, even when the setup was not completely rigid, the stress development was high enough for an effect on the strength of the cement to be expected (Bowen et al., 1983). In previous experiments (Hegdahl and Gjerdet, 1977; Davidson and De Gee, 1984), no effects of curing stress on the tensile strength of restorative composites could be demonstrated. The explanation was that flow would compensate completely for the stress development. Due to the unfavorable configuration of the system in the present study, the possibility of flow from the outside free surface can almost be excluded. Moreover, the viscosity of the cement is so high and the polymerization period so short, that flow is highly restrained. It has to be noted that this type of cement contains
voids as a result of being mixed. Growth of the voids due to the polymerization shrinkage stress contributes positively to the free surface and thus to the configuration. Consequently, the presence of voids (porosity) increases the flow capacity of a cement. In this study, only a slight pressure was exerted until the pre-adjusted cement layer thickness was reached. When porcelain or composite inlays are cemented clinically, only low pressure can be exerted, thus lowering the risk of premature fracture of the brittle inlay. Due to the relatively poor seat (approx. 100 plm) of the inlays, the cement layer under a clinical restoration will vary widely in thickness, leading to stress variations throughout the whole cement film. For adhesive luting of a facing, stress relief is drawn not only from the axial mobility of the restoration but also from the yielding of the restoration itself. The geometry of inlays, however, does not allow yielding. Therefore, our rigid set-up represents an example of the extreme situation that can occur under clinical conditions. Based on the preceding results and discussion, the following conclusions can be drawn: In both luting cements, tensile stress levels developed that exceeded the cohesive strengths of the materials. The effect of the thickness of the cement layer on the stress depends on the nature of the cement. The thicker the glass-ionomer cement layer, the faster the stress development. The composite cement showed an inverse relation. The contraction stress has a detrimental -effect on the cohesive strength of the glass-ionomer cement, but not on the cohesive strength of the composite lute in the present set-up. If microcracks are formed in the glass-ionomer cement, they will certainly contribute to premature failure of the cement. The glass-ionomer cement is capable of establishing a bond with the dentin substrate before the stress development starts, but the composite is not. This may lead to incomplete bonding and sealing by the composite lute. The difference in effect of contraction stress on the strength and the difference in establishing early bond strength to dentin may lead to different clinical effects of the glass ionomer and the composite applied as luting cements in crown and bridgework. In case of failure, one would prefer the mechanism of glass-ionomer cements, because fractures in (within) the bulk of the lute may still leave the dentin protected against bacterial invasion. This aspect of the different behavior patterns of glass-ionomer and composite lutes deserves further investigation.
Acknowledgment. We acknowledge the valuable support of the late Prof. dr. M. de Clercq during this study. REFERENCES BAUSCH, J.R.; DE LANGE, C.; DAVIDSON, C.L.; PETERS, A.; and DE GEE, A.J. (1982): Clinical Significance of Polymerization Shrinkage of Composite Resins, J Prosthet Dent
48:59-67.
BOWEN, R.L.; NEMOTO, K.; and RAPSON, J.E. (1983): Adhesive Bonding of Various Materials to Hard Tooth Tissues: Forces Developing in Composite Materials During Hardening, J Am Dent Assoc 106:475-477. DAVIDSON, C.L. and DE GEE, A.J. (1984): Relaxation of Polymerization Contraction Stresses by Flow in Dental Composites, J Dent Res 63:146-148. DAVIDSON, C.L.; DE GEE, A.J.; and FEILZER, A.J. (1984): The Competition between the Composite-Dentin Bond Strength and the Polymerization Contraction Stress, J Dent Res 63:1396-1399. FEILZER, A.J.; DE GEE, A.J.; and DAVIDSON, C.L. (1987): Setting Stress in Composite Resin in Relation to Configuration of the Restoration, J Dent Res 66:1636-1639. FEILZER, A.J.; DE GEE, A.J.; and DAVIDSON, C.L. (1989): Increased Wall-to-Wall Curing Contraction in Thin Bonded Resin Layers, J Dent Res 68:48-50. HEGDAHL, T. and GJERDET, N.R. (1977): Contraction Stresses of Composite Resin Filling Materials, Acta Odontol Scand 35:191-195. VAN ZEGHBROECK, L. (1989): Bond Capacity of Adhesive Luting Cements, Thesis, University of Louvain, Belgium, pp. 99-121. VAN ZEGHBROECK, L.; DAVIDSON, C.L.; and DE CLERCQ, M. (1989): Cohesive Failure due to Contraction Stress in Glass lonomer Luting Cements, J Dent Res 68:1014, Abst. No. 1180.
Downloaded from jdr.sagepub.com at UNIV OF MICHIGAN on June 17, 2015 For personal use only. No other uses without permission.
In this study, curing shrinkage stress development was monitored in a glass-ionomer and a BisGMA composite luting cement adhesively placed at film thicknesses ranging from 30 to 200 ALm. The nature and magnitude of the stress development depended greatly on the formulation and film thicknesses of the lute. The thicker the layer, the faster the stress development in the glass ionomer and the slower in the composite. The contraction stress had a detrimental effect on the cohesive strength of the glass ionomer and on the adhesive strength of the composite cement.
J Dent Res 70(5):880-882, May, 1991
Introduction. When the contraction of adhesive materials is obstructed, stresses are induced which can be detrimental for the cohesion of the structure (Bowen et al., 1983; Davidson et al., 1984). In particular, in cases with an unfavorable configuration-that is, where the ratio of bonded surface to free surface of the shrinking material is higher than five-all shrinkage is directed uniaxially, leading to a contraction that is almost three times higher than the expected linear shrinkage (Feilzer et al., 1989). Cement films will readily surpass such values. Moreover, tensile contraction stresses can exceed the adhesive or cohesive strength of the materials involved (Feilzer et al., 1987). Because all adhesive dental luting cements shrink considerably during setting, it is of interest to investigate the extent to which these cements fail when used as luting agents in rigid constructions. The purpose of the present study was to measure stress development in cement layers of various thicknesses and to demonstrate evidence of destruction of the lutes.
Materials and methods. One glass-ionomer cement [Fuji I-New, GC Int. Co., Japan, batch 130771(p) and 100771(1)] in a powder/liquid ratio of 1.8/1 and one BisGMA-based composite cement (Panavia Ex, Kuraray, Japan, batch no. 43121) in a powder/liquid ratio of 3.2/1 were tested. Stress development was determined at 24 PC in a setup as described by Feilzer et al. (1987), in which two opposing, identical, parallel stainless-steel disks (diameter = 10 mm) were connected to the load-cell and cross-head of a tensilometer (Instron Co., Bucks, UK). The axial sample contraction was continuously neutralized by a feedback displacement of the cross-head in order to maintain the original disk-to-disk distance with an accuracy of 0.2 pum. The steel disks were sandblasted with 250-ptm particles (Bego, Germany) and tin-plated (Kura-Ace tin plating, Kuraray Co., Ltd., Japan). This surface treatment ensured a bond strength that exceeded the cohesive strength of the materials under investigation. The cements were handled according to the manufacturer's instructions and were Received for publication June 13, 1990 Accepted for publication January 7, 1991 880
'Department of Prosthodontics,
inserted between the disks. Optimal adaptation was ensured by squeezing surplus material between the two disks until the desired disk-to-disk distance was obtained. Excess material was removed at the circumference of the disks. The glassionomer margin was coated with vaseline to prevent dehydration, and the composite margin was covered with Oxyguard® (as advised by the manufacturer) to promote complete setting. Two minutes from the start of mixing, continuous shrinkagestress recordings were made at pre-set disk-to-disk distances of 30, 50, 100, and 200 pum. At each film thickness, the average of five recordings was taken. So that the influence of the rigid set-up on the tensile strength of the cements could be determined, similar samples were tested at a cross-head speed of 0.5 mm/min at intervals of one min from four to 30 min after setting. Such experiments were done in a rigid set-up, as well as one in which the cements were allowed to shrink freely into the axial direction. Each experiment was repeated four times, and the results were averaged. The glass ionomer was tested at a film thickness of 30 pum, and the composite at 50 WLm, the minimum film thickness of the material (Van Zeghbroeck, 1989). In many cases, the samples broke spontaneously within the duration of the experiment. Non-linear regression analysis was carried out, and the significance of differences was determined at the 5% level by Student's t test. The measurement of the bond strength to dentin was carried out in the set-up which allowed free shrinkage into the axial direction. A series of dentin samples was cemented to a tincoated stainless-steel disk with the glass-ionomer luting cement in thicknesses of 30 ,um and with the composite in thicknesses of 50 ,um. The strength of the material was measured, at oneminute intervals, from four to 30 min after the start of the mixing.
All the fractured surfaces were examined under a stereomicroscope (X50) (Wild Heerbrugg, Switzerland) to check the nature of the fracture.
Results. Fig. 1 shows representative graphs of the shrinkage stress vs. time for each of the different glass-ionomer cement thicknesses. The nature of the stress development differed with the film thickness. The thicker the cement layer, the earlier the reaction started, and the faster and more continuous was the stress development. In thin layers (30 pum), the stress development showed a pronounced stick-slip contraction pattern. A representative graph of the stress development during setting of the composite is shown in Fig. 2. Unlike the stress development in the glass-ionomer cement, the reaction occurred in the composite cement faster in thinner layers, and the stress pattern was completely different. The setting reaction with this anaerobic setting material started immediately after application of the oxygen barrier and increased continuously until fracture. Non-linear regression analysis plots for the tensile strength as a function of time, for the glass ionomer in the free-shrinking and the rigid set-up, are given in Fig. 3. The correlation coefficients for both cases were, respectively, r2 =
Downloaded from jdr.sagepub.com at UNIV OF MICHIGAN on June 17, 2015 For personal use only. No other uses without permission.
STRESSES IN ADHESIVE LUTING CEMENTS
Vol. 70 No. 5
0.987 and r2 = 0.935 at p0.05). The
early tensile bond strengths of the glass-ionomer and the composite cement to dentin are shown in Figs. 3 and 4, respectively. The glass-ionomer cement always broke cohesively. The composite established its bond strength toward the dentin surface late, in comparison with the hardening stress development. The cement broke in a mixed cohesive/adhesive mode.
881
strength of the cement must develop more rapidly and be at all times superior to the tensile stresses that accompany polymerization. Within five min, the glass-ionomer luting cement had begun to establish a bond with dentin and the alloy. Shortly after gelation, when stress development starts (Bausch et al., 1982), the cement had reached its maximal bond strength to dentin (compare Figs. 1 and 3). Although the bond strength never exceeded 5 MPa, cohesive failure was observed in all cases. In the thinnest layers (30 ALm), the stress development grew gradually in time until a certain point, at which a sudden increase in stress development was registered, followed by the formation of a constant stress level during an interval period of time (see plateau parts in the curves). Although the relative error caused by the resolution of the measuring system increases with decreasing cement layer thickness, this does not explain the observed phenomena, because the resolution of the tensilometer in combination with the displacement measuring system was better than 0.04 jam. In addition, this phenomenon was not seen in the composite cement tested. An alternative explanation is that the glass ionomer forms microcracks due
Discussion.
20r FUJI-NEW 301um
In order to have a clinically useful adhesive bonding between the cement and the preparation and restoration, the bond
free I .m riaid
15
shrinkagege set-up set-un
L
-100 z
early bond streplength to dentin
ul
a:
5
A-- - --
-J
z
2-
FOE
U)
5
I
10
15
20
25
30 time (min)
LU
_
0)
-'._.- '
10
5
15
Fig.
20
25 30 time (min) . Fia 11-1Thp11rb< hrla-rci d IUlIUilUII U1f tLiIIl, fiuV 1-ilUW da; r'g6 iIaIdgr- btlase III IUJI. four different film thicknesses. The ends of the curves represent the points of cohesive failure.
ni..
20
3-The early tensile strength of Fuji I-New
as a
function of time
at a film thickness of 30 prm for a free-shrinkage and a rigid set-up. The shaded area indicates a 95% confidence interval. The development of the bond strength of the cement with dentin is indicated by the dotted line.
The bars
represent the S.D.
T
PANAVIA-EX
|,,,,X,~~ E
15
o1 0
10
E 0-
I I
C12 I I
'
a.
5 LU
5
I--
10
15
20
25
30
0)
5
10
15
20
25
30
time (min)
Fig. 2-The shrinkage stress in Panavia Ex as a function of setting time for three different film thicknesses. The dotted parts of the curves represent
Fig. 4-The early tensile strength of Panavia Ex as a function of time at a film thickness of 50 pum for a free-shrinkage and a rigid set-up. The shaded area indicates a 95% confidence interval. The development of the bond strength of the cement with dentin is indicated by the dotted line.
the range of cohesive failure.
The bars represent the S.D.
time (min)
Downloaded from jdr.sagepub.com at UNIV OF MICHIGAN on June 17, 2015 For personal use only. No other uses without permission.
882
DAVIDSON et al.
J Dent Res May 1991
E
earl -,V
crack formation
stag
,,,
=-rfl
curing contraction
Fig. 5-Schematic representation of crack formation and blunting of the crack tips due to prolonged tensile stress in a not-fully-set glass-ionomer cement. Cracks lead to cohesive failure in more brittle cement.
to the stress development in the incompletely set cement. Microcracks enlarge the free surface and relieve stress (Fig. 5). Due to the continuing setting process, the cement shrinks further and the stress build-up proceeds. However, when the cement ages, it becomes more brittle, so that newly-formed cracks cannot be blunted and stress relief is excluded, resulting in cohesive failure of the cement. The occurrence of plateaus in the stress-time curve had an inverse relation with the cement thickness. In the 200-pLm cement layers, plateaus were not
observed. This can be explained by a more homogeneous distribution of the curing shrinkage stresses in relatively thicker cement layers and the absence of the highly stressed regions that would normally induce microcracking. The abovementioned hypothesis of microcrack formation is supported by the fact that the strength of the cement is decreased by 25-30% in a rigid set-up. Moreover, the existence of microcracks in thin layers of glass ionomer, when cured in a rigid set-up, has been demonstrated microscopically (Van Zeghbroeck et al., 1989). This leads to the conclusion that the glass-ionomer cement is damaged by its own hardening stresses in a rigid set-up. The composite was able to establish a bond strength of almost 5 MPa to the dentin surface, but in contrast to the glassionomer cement, the bond with dentin was not yet formed when the hardening stress developed (Fig. 4). A cement with these hardening and setting characteristics is not likely to establish a clinical bond to the dentin cavity walls. The stress development showed an inverse relation with the film thickness of the composite (Fig. 2). This can be explained by the decreasing flow capacity of the cement with decreasing layer thicknesses. Due to a lower flow capacity in a thin layer, the shrinkage stress develops faster and thus surpasses sooner, and at a lower value, the cohesive strength of the cement (Feilzer et al., 1989). Once the oxygen barrier was placed on the composite, the stress developed continuously until fracture. Because the composite fractured cohesively in this set-up, the tensile strength of the cement could not be measured as a function of time in a completely rigid set-up with the two prepared steel disks. The stress development was allowed to reach only 85% of the total strength of the cement before the stressing was brought to a stop. However, even when the setup was not completely rigid, the stress development was high enough for an effect on the strength of the cement to be expected (Bowen et al., 1983). In previous experiments (Hegdahl and Gjerdet, 1977; Davidson and De Gee, 1984), no effects of curing stress on the tensile strength of restorative composites could be demonstrated. The explanation was that flow would compensate completely for the stress development. Due to the unfavorable configuration of the system in the present study, the possibility of flow from the outside free surface can almost be excluded. Moreover, the viscosity of the cement is so high and the polymerization period so short, that flow is highly restrained. It has to be noted that this type of cement contains
voids as a result of being mixed. Growth of the voids due to the polymerization shrinkage stress contributes positively to the free surface and thus to the configuration. Consequently, the presence of voids (porosity) increases the flow capacity of a cement. In this study, only a slight pressure was exerted until the pre-adjusted cement layer thickness was reached. When porcelain or composite inlays are cemented clinically, only low pressure can be exerted, thus lowering the risk of premature fracture of the brittle inlay. Due to the relatively poor seat (approx. 100 plm) of the inlays, the cement layer under a clinical restoration will vary widely in thickness, leading to stress variations throughout the whole cement film. For adhesive luting of a facing, stress relief is drawn not only from the axial mobility of the restoration but also from the yielding of the restoration itself. The geometry of inlays, however, does not allow yielding. Therefore, our rigid set-up represents an example of the extreme situation that can occur under clinical conditions. Based on the preceding results and discussion, the following conclusions can be drawn: In both luting cements, tensile stress levels developed that exceeded the cohesive strengths of the materials. The effect of the thickness of the cement layer on the stress depends on the nature of the cement. The thicker the glass-ionomer cement layer, the faster the stress development. The composite cement showed an inverse relation. The contraction stress has a detrimental -effect on the cohesive strength of the glass-ionomer cement, but not on the cohesive strength of the composite lute in the present set-up. If microcracks are formed in the glass-ionomer cement, they will certainly contribute to premature failure of the cement. The glass-ionomer cement is capable of establishing a bond with the dentin substrate before the stress development starts, but the composite is not. This may lead to incomplete bonding and sealing by the composite lute. The difference in effect of contraction stress on the strength and the difference in establishing early bond strength to dentin may lead to different clinical effects of the glass ionomer and the composite applied as luting cements in crown and bridgework. In case of failure, one would prefer the mechanism of glass-ionomer cements, because fractures in (within) the bulk of the lute may still leave the dentin protected against bacterial invasion. This aspect of the different behavior patterns of glass-ionomer and composite lutes deserves further investigation.
Acknowledgment. We acknowledge the valuable support of the late Prof. dr. M. de Clercq during this study. REFERENCES BAUSCH, J.R.; DE LANGE, C.; DAVIDSON, C.L.; PETERS, A.; and DE GEE, A.J. (1982): Clinical Significance of Polymerization Shrinkage of Composite Resins, J Prosthet Dent
48:59-67.
BOWEN, R.L.; NEMOTO, K.; and RAPSON, J.E. (1983): Adhesive Bonding of Various Materials to Hard Tooth Tissues: Forces Developing in Composite Materials During Hardening, J Am Dent Assoc 106:475-477. DAVIDSON, C.L. and DE GEE, A.J. (1984): Relaxation of Polymerization Contraction Stresses by Flow in Dental Composites, J Dent Res 63:146-148. DAVIDSON, C.L.; DE GEE, A.J.; and FEILZER, A.J. (1984): The Competition between the Composite-Dentin Bond Strength and the Polymerization Contraction Stress, J Dent Res 63:1396-1399. FEILZER, A.J.; DE GEE, A.J.; and DAVIDSON, C.L. (1987): Setting Stress in Composite Resin in Relation to Configuration of the Restoration, J Dent Res 66:1636-1639. FEILZER, A.J.; DE GEE, A.J.; and DAVIDSON, C.L. (1989): Increased Wall-to-Wall Curing Contraction in Thin Bonded Resin Layers, J Dent Res 68:48-50. HEGDAHL, T. and GJERDET, N.R. (1977): Contraction Stresses of Composite Resin Filling Materials, Acta Odontol Scand 35:191-195. VAN ZEGHBROECK, L. (1989): Bond Capacity of Adhesive Luting Cements, Thesis, University of Louvain, Belgium, pp. 99-121. VAN ZEGHBROECK, L.; DAVIDSON, C.L.; and DE CLERCQ, M. (1989): Cohesive Failure due to Contraction Stress in Glass lonomer Luting Cements, J Dent Res 68:1014, Abst. No. 1180.
Downloaded from jdr.sagepub.com at UNIV OF MICHIGAN on June 17, 2015 For personal use only. No other uses without permission.
