Solvent Degradation and Reduced Fracture Toughness in Aged Composites J.L. FERRACANE and V.A. MARKER' Department of Dental Materials Science, Oregon Health Sciences University, 611 S.W. Campus Drive, Portland, Oregon 97201; and 'Department of Dental Materials, Baylor College of Dentistry, 3302 Gaston Avenue, Dallas, Texas 75246 Quartz- and barium-glass-filled composites aged for more than one year in ethanol experienced a significant reduction in fracture toughness (Kk), essentially identical to that experienced after two months of aging. This reduction is mainly attributed to a softening of the resin matrix, but cracking within the resin and at the filler! matrix interface, as revealed by SEM microscopy, may also have contributed. No significant cracking could be seen in the composites aged in water. Composites post-cured at temperatures approaching their glass-transition temperature also experienced a reduction in after alcohol storage. Storage in water for one year had little NC effect on the K, of composites cured at oral temperatures, but a significant increase was observed for those post-cured at elevated temperatures. This increase is difficult to explain, but appears to involve a filler/matrix interfacial phenomenon, because it was not observed in the unfilled resin. The results ofthis study demonstrate that an alteration in the fracture resistance and some degradation of the filler/matrix interface, as has been observed clinically, occur after long-term exposure of dental composites to certain solvents used as food-simulating liquids.
Reductions in compressive yield strength or fracture toughness after storage for up to one year in water or in aqueous solutions containing salt or sucrose have been reported, the magnitude being dependent upon filler composition (Lloyd, 1984; Ferracane et al., 1987a). A significant increase in creep and decrease in flexural strength have also been demonstrated for certain composites stored in water for three months (Oysaed and Ruyter, 1986). Wu (1982) and Asmussen (1984) have demonstrated an appreciable softening and decrease in compressive strength for composites stored for ten days in solvents such as ethanol and acetone, as well as in organic acids which are present in plaque. Pilliar et al. (1986) demonstrated a decrease in fracture toughness (mini short-rod specimen design) for several composites stored for one or six months in water, and Ferracane et al. (1987a) reported a drastic reduction in fracture toughness (SEN bars) for various composites stored in ethanol for six months. Truong and Tyas (1988) reported a reduced fracture toughness (double torsion mode) for several types of composites after four to six weeks of storage in a 75% ethanol/water mixture, butthe effectwas material-dependent. In contrast, Pilliaretal. (1987) showed an increase in fracture toughness (mini short-rod design) for J Dent Res 71(l):13-19, January, 1992 several composites stored in ethanol for periods of up to one month. These disparities may be due to differences in the actual fractureIntroduction. toughness techniques used in the studies, the aging period, or compositional differences among the materials. Recently, Tyas Despite significant improvements in the formulation of modern (1990) demonstrated a correlation between fracture toughness, dental composites, the limited durability of these restoratives has elastic modulus, and tensile strength with surface chipping and restricted their applications. Microfill resins have exhibited a bulk fracture of class IV composites in vivo. relatively high incidence of chip fractures in anterior restorations It is apparent that some ofthe mechanical properties which may (Lambrechts et al., 1982; van Dijken et al., 1985; Tyas, 1990) and determine wear and fracture of composites can be compromised by brittle fracture in posterior restorations (Heymann et al., 1986). aging of the composites in various solvents which are used as foodSignificant wear has been reported in the occlusal contact areas, as simulating liquids (McKinney and Wu, 1985). Previous studies well as inthe contact-free areas ofposterior restorations (Lambrechts, suggest that the magnitude ofthe effect may be dependent upon the 1983; Lutz et al., 1984). A degradation of composites in the oral degree of conversion, as well as the resin and filler composition ofthe environment in the absence of loading or abrasive forces has also material. Therefore, the purpose ofthis study was to use experimental been reported by Roulet and Walti (1984) and van Groeningen et al. composite systems of known formulation to determine the effects of (1986). The mechanism of wear has been associated with solvent filler type and size, as well as post-cure temperature, on the fracture degradation of the surface and subsurface, in conjunction with toughness of composites aged for up to one year in both a good fracture ofthe resin matrix or the filler/matrix interface (Wu et al., solvent (75% EtOH/water) and a poor solvent (water). In addition, 1984; Hengchang et al., 1989). scanning electron microscopy was used to compare the fracture Subsurface degradation of composites in vivo has been demon- surfaces ofthe composites. Specifically, the presence ofmicrocracks strated by Wu et al. (1984) and, to a lesser extent, by Mair (1991). in the resin matrix and filler/matrix interfacial failure were invesThis phenomenon has been reproduced in vitro by the storing of tigated. composites in solvents, such as ethanol, which have solubility parameters (12-13 [cal/cm3]"2) similar to that ofthe resin component of composites and appear in FDA guidelines as food-simulating Materials and methods. liquids (Wu and McKinney, 1982; McKinney and Wu, 1985). This Four experimental composites, two containing large (L) filler parsoftening has been correlated with a reduced resistance to wear in ticles and two containing small (S) filler particles, and an unfilled vitro (Wu and McKinney, 1982). In addition, Wu (1982) has shown resin were examined (Table 1). The resin (U), which composed 20that increasing the degree of conversion in composites results in 30 wt% of the composites, was a light-activated dimethacrylate lower solubility and less softening in solvents. copolymer consisting of: 49.3 wt% Bis-GMA, 49.3 wt% TEGDMA, The mechanical properties of various composites have been 0.8 wt% camphorquinone, 0.5 wt% dimethylaminoethyl methacrytested after long-term storage in solutions. The effect of such aging late, and 0.1 wt% butylated hydroxytoluene. Filler levels were appears to be dependent upon composition, as well as testing mode. dictated, to a certain extent, by our ability to disperse the particles in the given quantity of resin. Due to the higher surface-area-toReceived for publication March 12, 1991 volume ratio for the small-particle composites, lower filler levels Accepted for publication September 9, 1991 This investigation was supported in part by USPHS Research Grant DE were achieved, though they all remained in the range of those found 07079-06 from the National Institute ofDental Research, National Institutes in present commercial systems. The appropriate amounts of the fillers were mixed into the resin with a spatula in a darkened room. of Health, Bethesda, MD 20892. 13
J Dent Res
FERRACANE & MARKER
14
CURED 900C - AGED EtOH/H20
CURED 370C - AGED EtOH/H20 2.10
2.10
:+T
1.80 N
icE
January 1992
1.50
1.80 cm
1.50 E
1.20
1.20 0.
0.90
0.90
CL
IL..
0.60
A-
A
0.60
-
*~~~
A
0.30
0.30
l .~ l ~~~ l.
0.00
0.00
0.1
10
1
100
0.1
1000
10
Fig. 1-Fracture toughness (FT) vs. time (days) for the specimens cured at 370C and aged in ethanol/water: Unfilled resin = A; LB = +; LQ = A; SQ 0;
SB =0.
Fig. 2-Fracture toughness (FT) vs. time (days) for the specimens cured at 900C and aged in ethanol/water: Unfilled resin = A; LB = +; LQ = A; SQ =0;
CURED 370C
-
1000
TIME (days)
TIME (days)
=
100
SB =0.
CURED 900C
AGED WATER
-
AGED WATER
2.10
2.10
1.80
1.80 N
1.50
1.50 .-ft
E
1.20
T-
0.90
0.90
0.60
0.60
0.30
0.30
~~~
A
0.00
0.00 0.1
1
10
100
0.1
1000
Fig. 3-Fracture toughness (FT) vs. time (days) for the specimens cured at 370C and aged in water: Unfilled resin = A; LB = +; LQ = A; SQ = 0; SB 0.
Immediately prior to specimen formation, the pastes were re-mixed to avoid settling ofthe filler and then exposed to a vacuum (0.1 mm Hg) for 15 min for removal of air bubbles. Bar-shaped specimens (17 mm x 3 mm x 1.5 mm) of the singleedge notch geometry were produced in a split steel mold containing a razor blade insert at mid-span for production of a notch with a crack-to-width (a/w) ratio of 0.5. Each specimen was cured by illumination with a visible-light source (Prisma-lite, L.D. Caulk, Milford, DE) for a total of 120 s. First, the top surface of the bar was illuminated with three overlapping exposures of 30 s each. Then, the specimen was removed from the die, and the bottom surface was illuminated with three overlapping tensecond exposures. No pre-crack was generated from the notch tip prior to specimen-testing, due to considerations of time, cost, and difficulty in production of a sufficient number of specimens. A previous study revealed that the fracture-toughness values obtained by the testing of notched-only bars ofthis size may be slightly higher than those obtained from pre-cracked bars (Ferracane et al., 1987b).
1
10
100
1000
TIME (days)
TIME (days)
=
1.20
0.
Fig. 4-Fracture toughness (FT) vs. time (days) for the specimens cured at 900C and aged in water: Unfilled resin = A; LB = +; LQ = A; SQ =0 ; SB =
0.
Following the initial light-curing, specimens were either immediately placed into a vial of distilled water or 75% ethanol/water in an oven at 37TC, or subjected to a post-cure heat treatment of 24 h in distilled water in an oven at 90TC before being placed into the vials containing distilled water or 75% ethanol/water in the oven at 37TC. The temperature of 900C was chosen for the post-cure treatment, because it is in the region ofthe glass-transition temperature for this resin system (Ferracane and Greener, 1986). Infrared analysis by the first author has shown that such a post-curing heat treatment produces approximately a 10% increase in degree of conversion (75% vs. 65%) for this resin, as compared with postcuring at 37TC. The specimens were then aged at 370C in sealed vials containing either the distilled water or the 75% ethanol/waterfor periods of one day (baseline), two months (adequate time for equilibration of the specimen with the solution), and 14 ± 1.5 months (sufficient time for degradation). The solution in the vials was not exchanged during aging.
FRACPLRE IOU&HNESS OF AGED COMPOSIIS
Vo. 71 N. 1
15
imatrx (bar = 10 n), (b) Fig. 5-a) SEM micrograph ofan LQ specimen aged for 14 months in ethanol after, showing secondary cracking in the ren SEM micrograph of an LQ specimen aged for 14 months in ethanol/water' showing filler/matrix interfacial failure (bar = 10 ui. toughness (K ) was evaluated in three-point bending a Universal testing cross-head speed of 0O.127 mm/min machine (Instron Model 1125, Canton, MA) For the specimens Fracture
at
on
a
determined, a small sliver (approximately 500 pm long and 10-30 mfthici) ofe1nposite was shaved from the surface each specimen with a scalpel blade under a stereomicroscope. The sliver was then
aged for 14 months, a wax reservoir was fashioned around the lower placed onto a KCl window and analyzed in transmission (100 scans) beam support and was filled with the storage solvent, thns allowing the notched area of the sample to remain wet throughout the fracture process. However, since the notch area does not desiccate in the time required for the tests, tbe overall imporance of this
through an aperture of 50 W2 under a nitrogen gas purge One sliver ?romn eaci of two different specimens analyzed in this was
manner so that comparative values of DC could be provided for each or less between the two values), seres vacationn was
procedure is probably negligible. The fracture toughness, N(KMPa calculated according to the following equation. ml2), Rests. W 0xf(a/ The IR evaluation revealed a substantially lower DC for the small Ne [(P S particle bariumglass composite (34%) in comparison with the other where P peak load (N). S = span (0.01 m.) B specimen thiickess materials (62-68%). There was no diftererce in DC between spepci (in), W specimen width (in), and f(a/w) is calculated according to means aged in water and those aged in ethanol the equation provided in ASTM Standard E399. There was no difference in the initial K values obtained after At least five bars were tested for each experimental condition. specimens were aged in water for one day fbrcLQ, LB, and SQ (Table Analysis of variance and Tukey's test for multiple comparisons 2) All three had higher K than SB and the unfilled resin. Similar between means were used at the level of p
Reductions in compressive yield strength or fracture toughness after storage for up to one year in water or in aqueous solutions containing salt or sucrose have been reported, the magnitude being dependent upon filler composition (Lloyd, 1984; Ferracane et al., 1987a). A significant increase in creep and decrease in flexural strength have also been demonstrated for certain composites stored in water for three months (Oysaed and Ruyter, 1986). Wu (1982) and Asmussen (1984) have demonstrated an appreciable softening and decrease in compressive strength for composites stored for ten days in solvents such as ethanol and acetone, as well as in organic acids which are present in plaque. Pilliar et al. (1986) demonstrated a decrease in fracture toughness (mini short-rod specimen design) for several composites stored for one or six months in water, and Ferracane et al. (1987a) reported a drastic reduction in fracture toughness (SEN bars) for various composites stored in ethanol for six months. Truong and Tyas (1988) reported a reduced fracture toughness (double torsion mode) for several types of composites after four to six weeks of storage in a 75% ethanol/water mixture, butthe effectwas material-dependent. In contrast, Pilliaretal. (1987) showed an increase in fracture toughness (mini short-rod design) for J Dent Res 71(l):13-19, January, 1992 several composites stored in ethanol for periods of up to one month. These disparities may be due to differences in the actual fractureIntroduction. toughness techniques used in the studies, the aging period, or compositional differences among the materials. Recently, Tyas Despite significant improvements in the formulation of modern (1990) demonstrated a correlation between fracture toughness, dental composites, the limited durability of these restoratives has elastic modulus, and tensile strength with surface chipping and restricted their applications. Microfill resins have exhibited a bulk fracture of class IV composites in vivo. relatively high incidence of chip fractures in anterior restorations It is apparent that some ofthe mechanical properties which may (Lambrechts et al., 1982; van Dijken et al., 1985; Tyas, 1990) and determine wear and fracture of composites can be compromised by brittle fracture in posterior restorations (Heymann et al., 1986). aging of the composites in various solvents which are used as foodSignificant wear has been reported in the occlusal contact areas, as simulating liquids (McKinney and Wu, 1985). Previous studies well as inthe contact-free areas ofposterior restorations (Lambrechts, suggest that the magnitude ofthe effect may be dependent upon the 1983; Lutz et al., 1984). A degradation of composites in the oral degree of conversion, as well as the resin and filler composition ofthe environment in the absence of loading or abrasive forces has also material. Therefore, the purpose ofthis study was to use experimental been reported by Roulet and Walti (1984) and van Groeningen et al. composite systems of known formulation to determine the effects of (1986). The mechanism of wear has been associated with solvent filler type and size, as well as post-cure temperature, on the fracture degradation of the surface and subsurface, in conjunction with toughness of composites aged for up to one year in both a good fracture ofthe resin matrix or the filler/matrix interface (Wu et al., solvent (75% EtOH/water) and a poor solvent (water). In addition, 1984; Hengchang et al., 1989). scanning electron microscopy was used to compare the fracture Subsurface degradation of composites in vivo has been demon- surfaces ofthe composites. Specifically, the presence ofmicrocracks strated by Wu et al. (1984) and, to a lesser extent, by Mair (1991). in the resin matrix and filler/matrix interfacial failure were invesThis phenomenon has been reproduced in vitro by the storing of tigated. composites in solvents, such as ethanol, which have solubility parameters (12-13 [cal/cm3]"2) similar to that ofthe resin component of composites and appear in FDA guidelines as food-simulating Materials and methods. liquids (Wu and McKinney, 1982; McKinney and Wu, 1985). This Four experimental composites, two containing large (L) filler parsoftening has been correlated with a reduced resistance to wear in ticles and two containing small (S) filler particles, and an unfilled vitro (Wu and McKinney, 1982). In addition, Wu (1982) has shown resin were examined (Table 1). The resin (U), which composed 20that increasing the degree of conversion in composites results in 30 wt% of the composites, was a light-activated dimethacrylate lower solubility and less softening in solvents. copolymer consisting of: 49.3 wt% Bis-GMA, 49.3 wt% TEGDMA, The mechanical properties of various composites have been 0.8 wt% camphorquinone, 0.5 wt% dimethylaminoethyl methacrytested after long-term storage in solutions. The effect of such aging late, and 0.1 wt% butylated hydroxytoluene. Filler levels were appears to be dependent upon composition, as well as testing mode. dictated, to a certain extent, by our ability to disperse the particles in the given quantity of resin. Due to the higher surface-area-toReceived for publication March 12, 1991 volume ratio for the small-particle composites, lower filler levels Accepted for publication September 9, 1991 This investigation was supported in part by USPHS Research Grant DE were achieved, though they all remained in the range of those found 07079-06 from the National Institute ofDental Research, National Institutes in present commercial systems. The appropriate amounts of the fillers were mixed into the resin with a spatula in a darkened room. of Health, Bethesda, MD 20892. 13
J Dent Res
FERRACANE & MARKER
14
CURED 900C - AGED EtOH/H20
CURED 370C - AGED EtOH/H20 2.10
2.10
:+T
1.80 N
icE
January 1992
1.50
1.80 cm
1.50 E
1.20
1.20 0.
0.90
0.90
CL
IL..
0.60
A-
A
0.60
-
*~~~
A
0.30
0.30
l .~ l ~~~ l.
0.00
0.00
0.1
10
1
100
0.1
1000
10
Fig. 1-Fracture toughness (FT) vs. time (days) for the specimens cured at 370C and aged in ethanol/water: Unfilled resin = A; LB = +; LQ = A; SQ 0;
SB =0.
Fig. 2-Fracture toughness (FT) vs. time (days) for the specimens cured at 900C and aged in ethanol/water: Unfilled resin = A; LB = +; LQ = A; SQ =0;
CURED 370C
-
1000
TIME (days)
TIME (days)
=
100
SB =0.
CURED 900C
AGED WATER
-
AGED WATER
2.10
2.10
1.80
1.80 N
1.50
1.50 .-ft
E
1.20
T-
0.90
0.90
0.60
0.60
0.30
0.30
~~~
A
0.00
0.00 0.1
1
10
100
0.1
1000
Fig. 3-Fracture toughness (FT) vs. time (days) for the specimens cured at 370C and aged in water: Unfilled resin = A; LB = +; LQ = A; SQ = 0; SB 0.
Immediately prior to specimen formation, the pastes were re-mixed to avoid settling ofthe filler and then exposed to a vacuum (0.1 mm Hg) for 15 min for removal of air bubbles. Bar-shaped specimens (17 mm x 3 mm x 1.5 mm) of the singleedge notch geometry were produced in a split steel mold containing a razor blade insert at mid-span for production of a notch with a crack-to-width (a/w) ratio of 0.5. Each specimen was cured by illumination with a visible-light source (Prisma-lite, L.D. Caulk, Milford, DE) for a total of 120 s. First, the top surface of the bar was illuminated with three overlapping exposures of 30 s each. Then, the specimen was removed from the die, and the bottom surface was illuminated with three overlapping tensecond exposures. No pre-crack was generated from the notch tip prior to specimen-testing, due to considerations of time, cost, and difficulty in production of a sufficient number of specimens. A previous study revealed that the fracture-toughness values obtained by the testing of notched-only bars ofthis size may be slightly higher than those obtained from pre-cracked bars (Ferracane et al., 1987b).
1
10
100
1000
TIME (days)
TIME (days)
=
1.20
0.
Fig. 4-Fracture toughness (FT) vs. time (days) for the specimens cured at 900C and aged in water: Unfilled resin = A; LB = +; LQ = A; SQ =0 ; SB =
0.
Following the initial light-curing, specimens were either immediately placed into a vial of distilled water or 75% ethanol/water in an oven at 37TC, or subjected to a post-cure heat treatment of 24 h in distilled water in an oven at 90TC before being placed into the vials containing distilled water or 75% ethanol/water in the oven at 37TC. The temperature of 900C was chosen for the post-cure treatment, because it is in the region ofthe glass-transition temperature for this resin system (Ferracane and Greener, 1986). Infrared analysis by the first author has shown that such a post-curing heat treatment produces approximately a 10% increase in degree of conversion (75% vs. 65%) for this resin, as compared with postcuring at 37TC. The specimens were then aged at 370C in sealed vials containing either the distilled water or the 75% ethanol/waterfor periods of one day (baseline), two months (adequate time for equilibration of the specimen with the solution), and 14 ± 1.5 months (sufficient time for degradation). The solution in the vials was not exchanged during aging.
FRACPLRE IOU&HNESS OF AGED COMPOSIIS
Vo. 71 N. 1
15
imatrx (bar = 10 n), (b) Fig. 5-a) SEM micrograph ofan LQ specimen aged for 14 months in ethanol after, showing secondary cracking in the ren SEM micrograph of an LQ specimen aged for 14 months in ethanol/water' showing filler/matrix interfacial failure (bar = 10 ui. toughness (K ) was evaluated in three-point bending a Universal testing cross-head speed of 0O.127 mm/min machine (Instron Model 1125, Canton, MA) For the specimens Fracture
at
on
a
determined, a small sliver (approximately 500 pm long and 10-30 mfthici) ofe1nposite was shaved from the surface each specimen with a scalpel blade under a stereomicroscope. The sliver was then
aged for 14 months, a wax reservoir was fashioned around the lower placed onto a KCl window and analyzed in transmission (100 scans) beam support and was filled with the storage solvent, thns allowing the notched area of the sample to remain wet throughout the fracture process. However, since the notch area does not desiccate in the time required for the tests, tbe overall imporance of this
through an aperture of 50 W2 under a nitrogen gas purge One sliver ?romn eaci of two different specimens analyzed in this was
manner so that comparative values of DC could be provided for each or less between the two values), seres vacationn was
procedure is probably negligible. The fracture toughness, N(KMPa calculated according to the following equation. ml2), Rests. W 0xf(a/ The IR evaluation revealed a substantially lower DC for the small Ne [(P S particle bariumglass composite (34%) in comparison with the other where P peak load (N). S = span (0.01 m.) B specimen thiickess materials (62-68%). There was no diftererce in DC between spepci (in), W specimen width (in), and f(a/w) is calculated according to means aged in water and those aged in ethanol the equation provided in ASTM Standard E399. There was no difference in the initial K values obtained after At least five bars were tested for each experimental condition. specimens were aged in water for one day fbrcLQ, LB, and SQ (Table Analysis of variance and Tukey's test for multiple comparisons 2) All three had higher K than SB and the unfilled resin. Similar between means were used at the level of p
