The Dependence of Shrinkage Stress Reduction on Porosity Concentration in Thin Resin Layers D. ALSTER, A.J. FEILZER, A.J. DE GEE, A. MOL1, and C.L. DAVIDSON Department of Dental Materials Science and 'Department of Oral Radiology, ACTA, Louwesweg 1, 1066 EA Amsterdam, The Netherlands
The development of polymerization contraction stress was determined as a function of the surface area of porosity, so that the contribution of voids in resin composite to stress relief could be investigated. Experiments were carried out on 200-mm-thick layers ofresin bonded from wall to wall in a restrained condition. The resin samples were divided into three groups: Group A was without porosity, group B contained a small number of pores, and group C contained a large number of pores in comparison with group B. For each group, porosity area, maximal stress, and stress development rate were determined. The mean maximal stress and stress development rate were inversely proportional to the mean porosity surface. These characteristics differed significantly (p < 0.01) between group A and C. For determination of whether shrinkage stress reduction has to be ascribed to flow from the outer surfaces of the voids or to inhibition of the setting reaction by oxygen in the voids, resin containing only nitrogen bubbles was also tested. The results indicated that both aspects contributed substantially to shrinkage stress relief. Incorporation of pores by the stirring of a luting cement contributes to stress reduction and can therefore be considered as a contribution to the maintenance of marginal integrity. J Dent Res 71(9):1619-1622, September, 1992
Introduction. Reduction of problems related to polymerization shrinkage is considered to be a main advantage of the inlay technique over bulk resin restorations. Attempts to minimize the polymerization shrinkage may be based on the controversial idea that luting cements in thin films will produce only a negligible shrinkage stress. However, Feilzer et al. (1987) have demonstrated that high stress develops in a resin cement film as a consequence of its extremely unfavorable ratio of bonded to outer unbonded surface. Indeed, besides a relatively high percentage of continuous margins, openings between the luting cement and dentin are frequently reported in the literature and ascribed to shrinkage stress (Douglas et al. 1989; Hflrzeler et al., 1990; Scherer et al., 1990; Krejci et al., 1992). Yet, a location of the total outline in enamel does not necessarily guarantee a perfect marginal integrity. In the occlusal areas of such restorations, the results are often satisfactory, but marginal gap formation and microleakage are reported in the cervical areas of MOD boxes and veneer restorations (Sorensen et al., 1990; Hannig et al., 1991; Lutz et al., 1991). The explanation of this may be sought in the lower yielding ability of the supporting cervical tooth structure. Lutz et al. (1991) demonstrated that cuspal movement is able to cancel out shrinkage stress of the cement layer. In studies where a light-cured as well as a dual-cured cement was tested, the latter showed a better marginal adaptation (Krejci et al., 1990; Noack et al.,1990; Hannig et al., 1991). It was hypothesized that admixing porosity contributes to shrinkage stress, since porosity enlarges the free surface area of the Received for publication January 6, 1992 Accepted for publication April 2, 1992
restoration and thus enlarges flow capacity (Davidson et al., 1991). The aim of the present study was to investigate porosity quantitatively as a potential source for stress reduction under restrained conditions of shrinkage in bonded thin unfilled resin layers. For investigation of the role of oxygen inhibition, the experiment was carried out under two conditions: with and without oxygen.
Materials and methods. So that the incorporation and determination of voids could be facilitated, an experimental unfilled resin for light-cured posterior restorations (EXI Lot 33, 3M Co., St. Paul, MN) was tested. A metal capsule, which could be well-sealed, was filled with a fixed amount of resin and two small metal balls and mixed for 30 s in a Silamat mixer (Vivadent, Liechtenstein). In this way, foam-like mixes were produced. Mixes containing less porosity were achieved by dilution of this foam-like mix with unmixed resin and manual stirring. One part of each mix was used for determination of the amount of porosity, and the other part was used for determination of setting stress development. The amount of porosity was determined in slabs of material that had been cured between two glass plates at a thickness of 200 mm. With a camera mounted on a light microscope (Carl Zeiss, Germany; magnification 50 X), slides were made of these samples with transmitted light. The total (spherical) inner surfaces of the pores were calculated on digitized images of these slides, with use of a PC-based image-analysis system. The developing stress in 200-pm-thick layers of polymerizing resin bonded from wall to wall was measured with the device described by Feilzer et al. (1987). The experimental set-up consisted of two opposing identical steel disks, covered with glass disks of equal diameter (0 = 10 mm and height = 2 mm). For the adhesion of a glass disc to the steel surface, the glass disc was roughened with SiC grinding powder (grit 320), treated with Scotchprime (3M Co.), and bonded to the steel surface with Gluma Sealer (Bayer Dental AG, Leverkusen, Germany). One steel disk was connected to the load-cell and the other to the cross-head of a tensilometer (Instron Co., Macclesfield, UK). So that the bond with the experimental resin layer would be enhanced, the glass disks were roughened with SiC grinding powder (grit 320) and were silane-coated (Scotchprime, 3M Co.). Portions of various mixes of resin were inserted between the glass disks and within 2 min after being mixed were cured by light (Visilux 2, 3M Co.) for 1 min from two sides. The axial sample contraction was continuously counteracted by a feed-back displacement of the cross-head so that the original disk-to-disk distance of 200 mm would be maintained. Simultaneously, recordings of setting stress development in time were made. The increase in stress during the first minute of stress development was taken as a measure for the stress development rate. The value at which the stress-time curve reached a plateau was registered as maximal stress. For maximal stress value calculations, only the results of samples showing no adhesive or cohesive failure due to shrinkage stress were taken into account. 1619
1620
ALS7TR et al.
J Dent Res Setember 1992
Fig. I Tosi estnoin eim of a hand nixed luti ce ment Bniant Duo Ceren), illu a i the ihmoenety of porosity Porosity area/volumi 0.6mm m 3 2 Scalehars 200 pm.
resin is 2.0 ummm3 (1) a
M
=
grop the mean a~nd standard devaton of pore surf ce area values per volume of rsin, maximal stress values and stress development rates. For comparison ofthe relative contributions of porosit arnd sample externral fre surface area to the total free area of the sample iean and standard deviations of the ototl fporle surface area inclusive of the externall ftee surface 1 resn sample are also given in the a ca1 samples wi h dufrent levels of porosity Porosit in hand-mixed area" area of wac s of Table* The percentage off lures per group is listed as well. A uting composites was difficult to var since different miring produced about the same level of porosit. Moreover- one-way NOVA test revealed sigifiant difcbrences among the mixes obtained with an unfilled resin had a more homogeneous fbr grups f both stress (F = 6,86 df=31 1 p < 0 01) and stres distribution of pores. The transmission of curng ligt is easier development rate values (F 67890 df 3.24 p < 0.01) A Student in a translucent resin Finally the light-colored background on t test showed a sigrficant diffrence between groups A and C slides made with transmitted light of the light microscope pro- f r both stres and stress development rate vlues vided a better contrast with the black-colored pores, whi ficilitated the image analysis For convnence of comparisn Disusson. in Fig. I the posityimage of two stes of one mixed composite tis iven In studies that tested both dualcun'd and hligt-cured luting Iue cement better marginal adaptation was obtained with the dualr curin material (Krj i et al., 1990' Rannig et al., 1991) The Results. authors hypothesized hat shrinkage stress development was Fig. 2 shows various levels of porosity as seen by translucent lght favorably influenced by mixing-porosity Davidso arnd De Gee (1984) demonstraed that shrinkage microscopy through a slab of 200 pm thickness. Fig. 3 shows examples of stress development as a cnction of polymer izaton stress causes How of the free outer unbonded surface of a resin time or the groups A B, C, and D. The Table gives, fr each restorationr wvch compensates partially for shrinkage. Felzer The experiments (n 5) were carried out with three types of m1 x (group C3 mixes, ie., pore-ee resin (group A) a fam-like and a blend of he previous two miles groupp B). Moreover, experiments with a foam like mix produced under nritrogen at mosphere were also carried out (group D) (Fig 2). An unfilled, resin was used to Encilitate the preparation of
=
=
_
TABLE
MEAN POROSITY AREANOLUME RESIN, TOTAL SAMPLE UNBONDED AREA, MAXIMAL STRESS S per GROUP STRESS DEVELOPMENT RATE MPa/nin) AND PERCEN E OF FAIL Maximal stress Stress development Por. area! vol.resin Total sample unbonded area mMu % filure ratet (MPa)t Mear Mean Mea (MPa/in) Group (SD) (SD) (SD) group -0 80 A 9.2 (1-1) 220(1.4 B 71 0.38 (0.11) 12.29 (1 66) 18.3 (314) 3.4 (2.0) £ 0 1.4 (0.5) 33.35 (16.83) 12.8 (2.1) 1.14 (0.57) 0 D 11.6 (3.3) 54*71 (22.07) 0.8 (0.3) 3.08 (1.41) Calculated fioi porosity arew/vulume resin, sample voluie, and external free ara of the sample Oacket) t Calculated only 'om samples showing no adhesive or cohesive failure due to shriage sress. t Calculated from the part of the curve dung he first minute. § External free area of the sample (jaekt).
Vol. 71 No 9
1621 P STRESS REDUCTON IN RESlY BYPORSUY12
Fig. 2Represertatie examples oflevels of orosit as sen wit ransritted ught mcroscopy Porosi 0.35 imn mm3 (B), 1.67 mm mm3 (C), and 4.25 mi2Wmn3 (D3. Scale bars = 200 pi. et a. (1987, 1991) showed that the magnitude of the remaining setting stress in composites increases with increasing ratio of bonded tothe free unbonded surface area of a restoration. In resin used as a luting cement this free outer surface is relatively small and is not likely to provide enough stress reduction. However, porosity in a lting material due to the mixing of two pastes can also be considered as free surface area. In a normally mixed duaI cure cement (we took as an example Brilliant Duo Cement, Colte'ne AG Altstaten, Switzerland) the total free surface area from pores is approximately 1 mm2/mn (Fig I). For example, when the cement layer thickness is 50 pm ithe restoration margin length is 50 mm, and the inner surface area of an inla restoration is put at 100 mM2 the contribution of the pores to bee surface is approximately twice that of the outer free surface area of the cement layer. The true ratio between bonded anid unbonded surface is then 200% more favorable thanr the ratio tat is based only on the configuration (Feilzer et at., 198 7) The total free area might be sufficient to provide enough shrirkage stress reduction to prevent separation of the bond. The Table shows that an increasing porosit concentration (Fig. 2) in 200 pm-thick resin layers results in a decreasing shrinkage stress level and a decreasing rate of stress development, Examples of the stress curves for the four difrent stages of porosity concentration are shown in Fig 3 For practical reasons, the experiments were carried out with an unfilled tesin, but since only the resin is responsible for flow the results should be applicable for filled resins as well. For determination of whether only the voids are responsible
pheca revoun resir s negligibi A
for the efct4 or whether oxygen nhibition also plays a role, a fourth group groupp D) of samples with only nitrogencontaining oids was examined. These samples were prepared rn a manner similar to at used fr those in group C, but in an atmosphere of nitrogen. Yet the samples in group D appeared to contain nearly three times the :eee pore surface area of those in group C A tentative explanraion of this difference in porosity is that nitrogen is not soluble in he resin, while the oxygen part of air is v ry likely absorbed in the molecular structure. Based on the trends fr groups A-C group D should have the lowest stress However the mean stress value of group D was slightly below that ouf goup C and the mean stess developments rate was Cu Based on these results the followirg slightly hier n group cotclusionr wer d awr (a) Resin with oxygen-free voids (group D) provides shrinkage stress reduction (diffrence between groups A and D) (b) The relatively low stress reduction in group D, with respect to the trends fir groups A-C, is a result of a combined efct of stress relief by voids and oxygen. In the psent study, the height of the remaining stress and the rate of shrink stress development have been taken as a measure for flow capacity of pore-containing resin. Therefore efforts havc been made to ensure an adhesive bond to the opposing glass discs, strong enoiih to withstand the setting forces. So that other influens on stress reduction would be excluded, compliance of the measuring system was practically eliminated by a transducer controlled feed-back displacement of the ross-head During the experiment, the original distance of 200 Vm did not change A
ALSTER et al.
1622
J Dent Res September 1992
stress (MPa)
25
Acknowledgment. We thank Henk Groen for his help with statistical evaluation.
-
group A 20 -
/,
group B
REFERENCES
15-
group C
10
group D
Davidson CL, De Gee AJ (1984). Relaxation of polymerization contraction stresses by flow in dental composites. J Dent Res 63:146-148. Davidson CL, Van Zeghbroeck L, Feilzer AJ (1991). Destructive stresses in adhesive luting cements. J Dent Res 70:880-882. 5 Douglas WH, Fields RP, Fundingsland J (1989). A comparison between the microleakage of direct and indirect composite restorative sys0 tems. J Dent 17:184-188. 1 4 6 0 2 3 5 7 8 9 10 Feilzer AJ, De Gee AJ, Davidson CL (1987). Setting stress in composite time (min.) resin in relation to configuration of the restoration. J Dent Res 66:1636-1639. Fig. 3-Representative examples of setting stress development as a Feilzer AJ, De Gee AJ, Davidson CL (1991). Setting stresses in composfunction of polymerization time for the groups A, B, C, and D. ites for the two curing modes (abstract). J Dent Res 70:527. HannigM, Rahlf B, Schlichting B (1991). Das randverhalten von Kulzerconsequence of this non-compliance set-up was the high percentage kompositinlays unter simultaner mechanischer und thermischer offailures in the pore-free resin of group A (Table). Here, the stress belastung. Dtsch Zahnarztl Z 46:618-621. developed so rapidly that the still-immature adhesive bond failed in Hurzeler M, Zimmerman E, Mormann WH (1990). Marginale adaptation most cases. In the clinical situation, the degree of compliance varies von maschinell hergestellten onlays in vitro. Schweiz Monatsch with the location ofthe restoration in the tooth. Molar cusps are able Zahnmed 100:715-720. to cancel out the shrinkage stress in the occlusal area of an inlay Krejci I, Picco U, Lutz F (1990). Dentinhaftung bei zahnfarbenen adh restoration by moving toward each other (Lutz et al., 1991). In the siven MOD-sofortinlays aus komposit. Schweiz Monatschr Zahnmed cervical area, much less compliance is to be expected. This results 100:1151-1159. in a poorer marginal adaptation, even with margins ending in Krejci I, Sdgesser D, Lutz F (1992). Optimierung der dentinhaftung in enamel (Sorensen et al., 1990; Hannig et al., 1991). gemischtenklasse-V-kavitaten. SchweizMonatschrZahnmed 102:32In groups C and D, no failures ofthe bond occurred. This can be 37. ascribed to both the lower shrinkage stress level and the slower Lutz F, Krejci I, Barbakow F (1991). Quality and durability of marginal stress development. The lower stress level was caused by the adaptation in bonded composite restorations. DentMater 10:107-113. smaller effective shrinkage, which was in turn due to growth of the Noack MJ, Locke LS, RouletJF (1990). Marginal adaptation of porcelain pores. With regard to the slower stress development, both flow ofthe inlays luted with different composite materials (abstract). JDent Res resin material and inhibition of the setting reaction by oxygen 69:161. played a role. Scherer W, Caliskan F, Kaim J, Moss S, Vijayaraghavan T (1990). Our study has shown that porosity in resins leads to stress Comparison of microleakage between direct placement composites reduction and therefore may contribute to the maintenance of marand direct composite inlays. Gen Dent 38:209-211. ginal integrity of adhesively luted restorations. This may well be an Sorensen JA, Strutz JM, Avera SP, Materdomini D (1990). Marginal explanation for the better performance of chemically-curing composfidelity and microleakage of porcelain veneers fabricated by two ites (Krejcietal., 1990; Hannigetal., 1991). In this respect, presence techniques. Proceedings of a presentation before the Pacific Coast of some porosity in the material is notper se disadvantageous. Society of Prosthodontics, Napa (CA).
The development of polymerization contraction stress was determined as a function of the surface area of porosity, so that the contribution of voids in resin composite to stress relief could be investigated. Experiments were carried out on 200-mm-thick layers ofresin bonded from wall to wall in a restrained condition. The resin samples were divided into three groups: Group A was without porosity, group B contained a small number of pores, and group C contained a large number of pores in comparison with group B. For each group, porosity area, maximal stress, and stress development rate were determined. The mean maximal stress and stress development rate were inversely proportional to the mean porosity surface. These characteristics differed significantly (p < 0.01) between group A and C. For determination of whether shrinkage stress reduction has to be ascribed to flow from the outer surfaces of the voids or to inhibition of the setting reaction by oxygen in the voids, resin containing only nitrogen bubbles was also tested. The results indicated that both aspects contributed substantially to shrinkage stress relief. Incorporation of pores by the stirring of a luting cement contributes to stress reduction and can therefore be considered as a contribution to the maintenance of marginal integrity. J Dent Res 71(9):1619-1622, September, 1992
Introduction. Reduction of problems related to polymerization shrinkage is considered to be a main advantage of the inlay technique over bulk resin restorations. Attempts to minimize the polymerization shrinkage may be based on the controversial idea that luting cements in thin films will produce only a negligible shrinkage stress. However, Feilzer et al. (1987) have demonstrated that high stress develops in a resin cement film as a consequence of its extremely unfavorable ratio of bonded to outer unbonded surface. Indeed, besides a relatively high percentage of continuous margins, openings between the luting cement and dentin are frequently reported in the literature and ascribed to shrinkage stress (Douglas et al. 1989; Hflrzeler et al., 1990; Scherer et al., 1990; Krejci et al., 1992). Yet, a location of the total outline in enamel does not necessarily guarantee a perfect marginal integrity. In the occlusal areas of such restorations, the results are often satisfactory, but marginal gap formation and microleakage are reported in the cervical areas of MOD boxes and veneer restorations (Sorensen et al., 1990; Hannig et al., 1991; Lutz et al., 1991). The explanation of this may be sought in the lower yielding ability of the supporting cervical tooth structure. Lutz et al. (1991) demonstrated that cuspal movement is able to cancel out shrinkage stress of the cement layer. In studies where a light-cured as well as a dual-cured cement was tested, the latter showed a better marginal adaptation (Krejci et al., 1990; Noack et al.,1990; Hannig et al., 1991). It was hypothesized that admixing porosity contributes to shrinkage stress, since porosity enlarges the free surface area of the Received for publication January 6, 1992 Accepted for publication April 2, 1992
restoration and thus enlarges flow capacity (Davidson et al., 1991). The aim of the present study was to investigate porosity quantitatively as a potential source for stress reduction under restrained conditions of shrinkage in bonded thin unfilled resin layers. For investigation of the role of oxygen inhibition, the experiment was carried out under two conditions: with and without oxygen.
Materials and methods. So that the incorporation and determination of voids could be facilitated, an experimental unfilled resin for light-cured posterior restorations (EXI Lot 33, 3M Co., St. Paul, MN) was tested. A metal capsule, which could be well-sealed, was filled with a fixed amount of resin and two small metal balls and mixed for 30 s in a Silamat mixer (Vivadent, Liechtenstein). In this way, foam-like mixes were produced. Mixes containing less porosity were achieved by dilution of this foam-like mix with unmixed resin and manual stirring. One part of each mix was used for determination of the amount of porosity, and the other part was used for determination of setting stress development. The amount of porosity was determined in slabs of material that had been cured between two glass plates at a thickness of 200 mm. With a camera mounted on a light microscope (Carl Zeiss, Germany; magnification 50 X), slides were made of these samples with transmitted light. The total (spherical) inner surfaces of the pores were calculated on digitized images of these slides, with use of a PC-based image-analysis system. The developing stress in 200-pm-thick layers of polymerizing resin bonded from wall to wall was measured with the device described by Feilzer et al. (1987). The experimental set-up consisted of two opposing identical steel disks, covered with glass disks of equal diameter (0 = 10 mm and height = 2 mm). For the adhesion of a glass disc to the steel surface, the glass disc was roughened with SiC grinding powder (grit 320), treated with Scotchprime (3M Co.), and bonded to the steel surface with Gluma Sealer (Bayer Dental AG, Leverkusen, Germany). One steel disk was connected to the load-cell and the other to the cross-head of a tensilometer (Instron Co., Macclesfield, UK). So that the bond with the experimental resin layer would be enhanced, the glass disks were roughened with SiC grinding powder (grit 320) and were silane-coated (Scotchprime, 3M Co.). Portions of various mixes of resin were inserted between the glass disks and within 2 min after being mixed were cured by light (Visilux 2, 3M Co.) for 1 min from two sides. The axial sample contraction was continuously counteracted by a feed-back displacement of the cross-head so that the original disk-to-disk distance of 200 mm would be maintained. Simultaneously, recordings of setting stress development in time were made. The increase in stress during the first minute of stress development was taken as a measure for the stress development rate. The value at which the stress-time curve reached a plateau was registered as maximal stress. For maximal stress value calculations, only the results of samples showing no adhesive or cohesive failure due to shrinkage stress were taken into account. 1619
1620
ALS7TR et al.
J Dent Res Setember 1992
Fig. I Tosi estnoin eim of a hand nixed luti ce ment Bniant Duo Ceren), illu a i the ihmoenety of porosity Porosity area/volumi 0.6mm m 3 2 Scalehars 200 pm.
resin is 2.0 ummm3 (1) a
M
=
grop the mean a~nd standard devaton of pore surf ce area values per volume of rsin, maximal stress values and stress development rates. For comparison ofthe relative contributions of porosit arnd sample externral fre surface area to the total free area of the sample iean and standard deviations of the ototl fporle surface area inclusive of the externall ftee surface 1 resn sample are also given in the a ca1 samples wi h dufrent levels of porosity Porosit in hand-mixed area" area of wac s of Table* The percentage off lures per group is listed as well. A uting composites was difficult to var since different miring produced about the same level of porosit. Moreover- one-way NOVA test revealed sigifiant difcbrences among the mixes obtained with an unfilled resin had a more homogeneous fbr grups f both stress (F = 6,86 df=31 1 p < 0 01) and stres distribution of pores. The transmission of curng ligt is easier development rate values (F 67890 df 3.24 p < 0.01) A Student in a translucent resin Finally the light-colored background on t test showed a sigrficant diffrence between groups A and C slides made with transmitted light of the light microscope pro- f r both stres and stress development rate vlues vided a better contrast with the black-colored pores, whi ficilitated the image analysis For convnence of comparisn Disusson. in Fig. I the posityimage of two stes of one mixed composite tis iven In studies that tested both dualcun'd and hligt-cured luting Iue cement better marginal adaptation was obtained with the dualr curin material (Krj i et al., 1990' Rannig et al., 1991) The Results. authors hypothesized hat shrinkage stress development was Fig. 2 shows various levels of porosity as seen by translucent lght favorably influenced by mixing-porosity Davidso arnd De Gee (1984) demonstraed that shrinkage microscopy through a slab of 200 pm thickness. Fig. 3 shows examples of stress development as a cnction of polymer izaton stress causes How of the free outer unbonded surface of a resin time or the groups A B, C, and D. The Table gives, fr each restorationr wvch compensates partially for shrinkage. Felzer The experiments (n 5) were carried out with three types of m1 x (group C3 mixes, ie., pore-ee resin (group A) a fam-like and a blend of he previous two miles groupp B). Moreover, experiments with a foam like mix produced under nritrogen at mosphere were also carried out (group D) (Fig 2). An unfilled, resin was used to Encilitate the preparation of
=
=
_
TABLE
MEAN POROSITY AREANOLUME RESIN, TOTAL SAMPLE UNBONDED AREA, MAXIMAL STRESS S per GROUP STRESS DEVELOPMENT RATE MPa/nin) AND PERCEN E OF FAIL Maximal stress Stress development Por. area! vol.resin Total sample unbonded area mMu % filure ratet (MPa)t Mear Mean Mea (MPa/in) Group (SD) (SD) (SD) group -0 80 A 9.2 (1-1) 220(1.4 B 71 0.38 (0.11) 12.29 (1 66) 18.3 (314) 3.4 (2.0) £ 0 1.4 (0.5) 33.35 (16.83) 12.8 (2.1) 1.14 (0.57) 0 D 11.6 (3.3) 54*71 (22.07) 0.8 (0.3) 3.08 (1.41) Calculated fioi porosity arew/vulume resin, sample voluie, and external free ara of the sample Oacket) t Calculated only 'om samples showing no adhesive or cohesive failure due to shriage sress. t Calculated from the part of the curve dung he first minute. § External free area of the sample (jaekt).
Vol. 71 No 9
1621 P STRESS REDUCTON IN RESlY BYPORSUY12
Fig. 2Represertatie examples oflevels of orosit as sen wit ransritted ught mcroscopy Porosi 0.35 imn mm3 (B), 1.67 mm mm3 (C), and 4.25 mi2Wmn3 (D3. Scale bars = 200 pi. et a. (1987, 1991) showed that the magnitude of the remaining setting stress in composites increases with increasing ratio of bonded tothe free unbonded surface area of a restoration. In resin used as a luting cement this free outer surface is relatively small and is not likely to provide enough stress reduction. However, porosity in a lting material due to the mixing of two pastes can also be considered as free surface area. In a normally mixed duaI cure cement (we took as an example Brilliant Duo Cement, Colte'ne AG Altstaten, Switzerland) the total free surface area from pores is approximately 1 mm2/mn (Fig I). For example, when the cement layer thickness is 50 pm ithe restoration margin length is 50 mm, and the inner surface area of an inla restoration is put at 100 mM2 the contribution of the pores to bee surface is approximately twice that of the outer free surface area of the cement layer. The true ratio between bonded anid unbonded surface is then 200% more favorable thanr the ratio tat is based only on the configuration (Feilzer et at., 198 7) The total free area might be sufficient to provide enough shrirkage stress reduction to prevent separation of the bond. The Table shows that an increasing porosit concentration (Fig. 2) in 200 pm-thick resin layers results in a decreasing shrinkage stress level and a decreasing rate of stress development, Examples of the stress curves for the four difrent stages of porosity concentration are shown in Fig 3 For practical reasons, the experiments were carried out with an unfilled tesin, but since only the resin is responsible for flow the results should be applicable for filled resins as well. For determination of whether only the voids are responsible
pheca revoun resir s negligibi A
for the efct4 or whether oxygen nhibition also plays a role, a fourth group groupp D) of samples with only nitrogencontaining oids was examined. These samples were prepared rn a manner similar to at used fr those in group C, but in an atmosphere of nitrogen. Yet the samples in group D appeared to contain nearly three times the :eee pore surface area of those in group C A tentative explanraion of this difference in porosity is that nitrogen is not soluble in he resin, while the oxygen part of air is v ry likely absorbed in the molecular structure. Based on the trends fr groups A-C group D should have the lowest stress However the mean stress value of group D was slightly below that ouf goup C and the mean stess developments rate was Cu Based on these results the followirg slightly hier n group cotclusionr wer d awr (a) Resin with oxygen-free voids (group D) provides shrinkage stress reduction (diffrence between groups A and D) (b) The relatively low stress reduction in group D, with respect to the trends fir groups A-C, is a result of a combined efct of stress relief by voids and oxygen. In the psent study, the height of the remaining stress and the rate of shrink stress development have been taken as a measure for flow capacity of pore-containing resin. Therefore efforts havc been made to ensure an adhesive bond to the opposing glass discs, strong enoiih to withstand the setting forces. So that other influens on stress reduction would be excluded, compliance of the measuring system was practically eliminated by a transducer controlled feed-back displacement of the ross-head During the experiment, the original distance of 200 Vm did not change A
ALSTER et al.
1622
J Dent Res September 1992
stress (MPa)
25
Acknowledgment. We thank Henk Groen for his help with statistical evaluation.
-
group A 20 -
/,
group B
REFERENCES
15-
group C
10
group D
Davidson CL, De Gee AJ (1984). Relaxation of polymerization contraction stresses by flow in dental composites. J Dent Res 63:146-148. Davidson CL, Van Zeghbroeck L, Feilzer AJ (1991). Destructive stresses in adhesive luting cements. J Dent Res 70:880-882. 5 Douglas WH, Fields RP, Fundingsland J (1989). A comparison between the microleakage of direct and indirect composite restorative sys0 tems. J Dent 17:184-188. 1 4 6 0 2 3 5 7 8 9 10 Feilzer AJ, De Gee AJ, Davidson CL (1987). Setting stress in composite time (min.) resin in relation to configuration of the restoration. J Dent Res 66:1636-1639. Fig. 3-Representative examples of setting stress development as a Feilzer AJ, De Gee AJ, Davidson CL (1991). Setting stresses in composfunction of polymerization time for the groups A, B, C, and D. ites for the two curing modes (abstract). J Dent Res 70:527. HannigM, Rahlf B, Schlichting B (1991). Das randverhalten von Kulzerconsequence of this non-compliance set-up was the high percentage kompositinlays unter simultaner mechanischer und thermischer offailures in the pore-free resin of group A (Table). Here, the stress belastung. Dtsch Zahnarztl Z 46:618-621. developed so rapidly that the still-immature adhesive bond failed in Hurzeler M, Zimmerman E, Mormann WH (1990). Marginale adaptation most cases. In the clinical situation, the degree of compliance varies von maschinell hergestellten onlays in vitro. Schweiz Monatsch with the location ofthe restoration in the tooth. Molar cusps are able Zahnmed 100:715-720. to cancel out the shrinkage stress in the occlusal area of an inlay Krejci I, Picco U, Lutz F (1990). Dentinhaftung bei zahnfarbenen adh restoration by moving toward each other (Lutz et al., 1991). In the siven MOD-sofortinlays aus komposit. Schweiz Monatschr Zahnmed cervical area, much less compliance is to be expected. This results 100:1151-1159. in a poorer marginal adaptation, even with margins ending in Krejci I, Sdgesser D, Lutz F (1992). Optimierung der dentinhaftung in enamel (Sorensen et al., 1990; Hannig et al., 1991). gemischtenklasse-V-kavitaten. SchweizMonatschrZahnmed 102:32In groups C and D, no failures ofthe bond occurred. This can be 37. ascribed to both the lower shrinkage stress level and the slower Lutz F, Krejci I, Barbakow F (1991). Quality and durability of marginal stress development. The lower stress level was caused by the adaptation in bonded composite restorations. DentMater 10:107-113. smaller effective shrinkage, which was in turn due to growth of the Noack MJ, Locke LS, RouletJF (1990). Marginal adaptation of porcelain pores. With regard to the slower stress development, both flow ofthe inlays luted with different composite materials (abstract). JDent Res resin material and inhibition of the setting reaction by oxygen 69:161. played a role. Scherer W, Caliskan F, Kaim J, Moss S, Vijayaraghavan T (1990). Our study has shown that porosity in resins leads to stress Comparison of microleakage between direct placement composites reduction and therefore may contribute to the maintenance of marand direct composite inlays. Gen Dent 38:209-211. ginal integrity of adhesively luted restorations. This may well be an Sorensen JA, Strutz JM, Avera SP, Materdomini D (1990). Marginal explanation for the better performance of chemically-curing composfidelity and microleakage of porcelain veneers fabricated by two ites (Krejcietal., 1990; Hannigetal., 1991). In this respect, presence techniques. Proceedings of a presentation before the Pacific Coast of some porosity in the material is notper se disadvantageous. Society of Prosthodontics, Napa (CA).
