Journal of Oral Rehabilitation 2001 28; 1015±1021
Chemical degradation of composite restoratives
A . U . J . Y A P * , ², S . H . L . T A N * , S . S . C . W E E * , C . W . L E E * ,
E . L . C . L I M * & K . Y . Z E N G ² *Department of Restorative Dentistry, Faculty of Dentistry, and ²Institute of Materials Research
and Engineering, National University of Singapore, Singapore
SUMMARY The chemical environment is one aspect
of the oral environment, which could have an
appreciable in¯uence on the in vivo degradation of
composite restoratives. The effects of chemical
media on surface hardness of four composite restor-
atives (Silux [SX], Z100 [ZO], Ariston [AR] and
Sure®l [SF]) were investigated. The relationship
between hardness and the thickness of the degra-
dation layer was also studied. Thirty six specimens
(3 ´ 4 ´ 2 mm) were made for each material. Fol-
lowing polymerization, the specimens were stored
in arti®cial saliva at 37 °C for 24 h. The specimens
were then randomly divided into six groups of six,
subjected to microhardness testing (load = 500 gf,
dwell time = 15 s) and stored in the following
chemicals for 1 week at 37 °C: arti®cial saliva (S),
distilled water (W), 0á02 N citric acid (C), 0á02 N lactic
acid (L), heptane (H) and 75±25% ethanol±water
Introduction
A composite material may be de®ned as three-
dimensional combinations of at least two chemically
different materials with a distinct interface (Phillips,
1981). Dental composite resins essentially comprise of
the resin matrix (organic phase), ®ller-matrix coupling
agent (interface), ®ller particles (dispersed phase) and
other minor additions including polymerization initi-
ators, stabilizers and colouring pigments. From micro-
defect analysis on biopsies (Wu et al., 1984) containing
composite restorations, damage is observed on both
stress and non-stress bearing surfaces. Degradation of
composites in the oral environment in the absence of
loading and abrasive forces has also been reported by
Roulet and Walti (1984) and Van Groeningen et al.
(1986). The results stipulate that degradation of com-
ã 2001 Blackwell Science Ltd
solution (E). After conditioning, the specimens were
again subjected to hardness testing and sectioned.
Change in hardness (DH) was computed and the
thickness of the degradation layer (DL) was meas-
ured using a computerized image analysis system at
600´ magni®cation. Results of statistical analysis
(ANOVA /Scheffe's [P < 0á05]) of DH based on mater-
ials were as follows: SX ± E > all other mediums;
ZO ± W > C; and AR ± S, W, E > H (> indicates signi-
®cantly greater hardness change). No signi®cant
difference in DH was observed between the different
chemicals for SF. The effects of chemical media on
DH were found to be material dependent. A signi®-
cant but weak positive correlation (Pearson's corre-
lation [P < 0á05]) exists between change in hardness
and thickness of the degradation layer.
KEYWORDS: chemical, composite, degradation,
environment, hardness
posites cannot be attributed to wear alone, but involves
chemical degradation as well. In vivo, composite resins
may either be exposed intermittently or continuously
to chemical agents found in saliva, food and beverages.
Intermittent exposure occurs during eating or drinking
until teeth are cleaned. Continuous exposure may,
however, occur as chemical agents can be absorbed by
adherent debris (such as calculus or food particles) at
the margins of restorations or be produced by bacterial
decomposition of debris. Polymerization shrinkage and
the diffusion of moisture through the resin component
may lead to the initiation and propagation of micro-
cracks at the tooth±restoration interface and in the resin
matrix. This process could provide a supply of chemical
agents and a path for further diffusion into the
restorative material, which may result in more rapid
degradation (Lee et al., 1995).
1015
1016 A . U . J . Y A P et al.

Table 1. Technical pro®les of the composites evaluated

Filler size Filler content

Material Manufacturer Type Polymer Fillers (lm) (% by volume) Shade

Silux Plus 3M Dental Products, Micro®ll BisGMA Silica 0á04 40 Y

(Lot no. 19980106) St Paul, MN 55144 TEGDMA (mean)
Z100 3M Dental Products, Mini®ll BisGMA Zirconia 0á5±0á7 66 A2
(Lot no. 19980203) St. Paul, MN 55144 TEGDMA Silica (mean)
Sure®l Dentsply-Caulk, Mini®ll Urethane- Ba-Boron- 0á8 65 A
(Lot no. 980709) Milford, DE 19963 modi®ed Fluorosilicate (mean)
BisGMA glass, Silica
Ariston pHc Vivadent Schaan, Midi®ll BisGMA Ba-Al- 1á3 59 U
(Lot no. A06719) Liechtenstein Fluorosilicate (mean)
UDMA Glass
Dimeth-acrylate Alkaline, Glass
Silica, Ytterbium
Tri¯uoride

BisGMA ˆ Bisphenol A-glycidyl methacrylate.

TEDGMA ˆ Triethylene glycol dimethacrylate.
UDMA ˆ Urethane dimethacrylate.

The chemical environment is one aspect of the oral
environment, which could have an appreciable in¯u-
ence on the in vivo degradation of composite restor-
atives. This study investigated the effects of six chemical
media (arti®cial saliva, distilled water, citric acid, lactic
acid, heptane and ethanol±water solution) on the
surface hardness of four composite restoratives. The
thickness of the degradation layer (DL), which is
de®ned as the outermost layer of composite that is
affected by the action of chemicals, was also measured
and correlated to change in hardness (DH).
Materials and methods
The visible-light cured composite restoratives investi-
gated are listed in Table 1. The composites were placed
in rectangular recesses (4 mm length ´ 3 mm
breath ´ 2 mm depth) of customized acrylic moulds
and covered with acetate strips*. A glass slide was
placed over the acetate strips and pressure was applied
to extrude excess material. The restoratives were light-
polymerized according to manufacturers' cure-times
(40 s) through the glass slide with a Spectrum Curing
Light². Immediately after light polymerization, the
acetate strips were discarded and the composites were
stored in arti®cial saliva³ for 24 h at 37 °C. A total of 36
*Hawe-Neos Dental, Bioggio, Switzerland.
² ¶
Dentsply Inc., Milford, DE 19963, USA.
³
NUH Pharmacy Laboratory, Singapore.
specimens were made for each material. After the 24 h
storage period, the specimens blotted dry and subjected
to hardness testing. The specimens were positioned
centrally beneath the indenter of a digital microhard-
ness tester§ (FM7) and a 500 gf load was applied
through the indenter with a dwell time of 15 s. The
Knoop Hardness Number (KHN) of each specimen was
recorded (I1). The specimens were then randomly
divided into six groups of six specimens and condi-
tioned for 1 week at 37 °C in arti®cial saliva control (S)
and the following chemicals: distilled water (W), 0á02 N
citric acid (C), 0á02 N lactic acid (L), heptane (H) and
75±25% ethanol±water solution (E). After conditioning
in arti®cial saliva and the various food-simulating
liquids for 1 week at 37 °C, the hardness of individual
specimens were re-measured (I2) at a site adjacent to
the initial indentation. The change in hardness (DH)
was then calculated using the following equation:
DH ˆ I1 ÿ I2
Where I1 is the KHN pre-conditioning and I2 is the KHN
after conditioning.
After microhardness testing, the specimens were
sectioned breath-wise with a diamond disc at
3000 rpm. using a sectioning device (Accutom 5)¶.
The cross-sections of the specimens were then viewed at
600´ magni®cation using a stereomicroscope** (BX60)
§
Future-Tech Corp., Tokyo, Japan.
Struers, Copenhagen, Denmark.
**Olympus, Tokyo, Japan.

ã 2001 Blackwell Science Ltd, Journal of Oral Rehabilitation 28; 1015±1021

 CHEMICAL DEGRADATION OF COMPOSITE RESTORATIVES 1017

Table 2. Mean change in hardness

Silux Plus Z100 Ariston pHc Sure®l
(DH) and thickness of degradation
layer (DL) in lm for the different Material Medium DH DL DH DL DH DL DH DL
composites
Arti®cial saliva ±3á35 6á94 ±1á76 3á12 16á13 43á95 ±1á65 6á61
(3á45) (1á01) (16á40) (0á94) (11á54) (4á56) (6á85) (0á69)
Distilled water ±0á98 6á51 20á38 6á92 9á48 14á57 ±0á10 7á33
(1á81) (0á96) (12á91) (0á39) (6á27) (0á96) (4á13) (0á44)
Citric acid ±3á97 7á61 ±7á25 6á62 4á10 12á42 ±6á37 9á31
(2á01) (1á92) (12á91) (0á67) (3á96) (2á11) (5á99) (0á85)
Lactic acid ±3á22 10á38 2á93 5á32 5á78 14á90 11á02 9á33
(2á10) (0á94) (6á71) (0á57) (3á41) (1á94) (12á93) (0á85)
Heptane ±0á23 18á55 1á23 6á20 ±5á15 14á13 4á02 3á98
(2á79) (3á17) (11á82) (1á59) (2á78) (1á21) (16á73) (0á66)
Ethanol±water 9á40 10á51 4á56 7á55 17á27 6á54 8á65 4á81
Solution (4á86) (0á39) (9á32) (0á41) (7á12) (7á60) (10á97) (0á28)

Standard deviation (SD) in brackets.

– s

Fig. 1. Mean change in hardness − −

(DH) for the different composites.

and computerized image analysis software (analysis
3á0)²². The thickness of the degradation layer (DL) of
each specimen was obtained from the average of three
readings. The mean DH and DL for each material±
chemical combination was calculated and tabulated. For
all statistical analyses, a signi®cance level of 0á05 was
used. Two-way analysis of variance (ANOVA ) was
performed on DH and DL data, with restorative material
and chemical medium as main effects, and all possible
combinations of these variables as interaction effects in
the ANOVA model. Post hoc Scheffe's test was used to
test for differences among means³³. One-way ANOVA
was also performed with materials and chemical
medium as independent variables to determine the
²²
Softimaging Systems, Frankfurt, Germany.
³³
SPSS Inc., Chicago, IL 60611, USA.
effects of chemical media on DH/DL of individual
composites and compare DH/DL between different
composites in the same medium. Correlation DH and
DL was done using Pearson's product±moment correla-
tion at a signi®cance level of 0á01.
Results
The mean DH values and thickness of DL of the
different composites are shown in Table 2 and Figs 1
and 2. Results of statistical analysis are presented in
Tables 3 and 4. A negative DH, which indicates an
increase in hardness, was observed for Silux after
conditioning in all media with the exception of
ethanol±water solution. For Z100 and Ariston, an
increase in hardness was observed after conditioning
in arti®cial saliva and citric acid, and heptane,

ã 2001 Blackwell Science Ltd, Journal of Oral Rehabilitation 28; 1015±1021

1018 A . U . J . Y A P et al.

– s

s
Fig. 2. Mean thickness of degrada-
tion layer (DL) for the different
composites.

Table 3. Results of statistical

Materials Property Differences
analysis based on materials
Silux Plus DH Ethanol±water solution > saliva, distilled water, citric acid, lactic acid,
heptane
DL Heptane > saliva, distilled water, citric acid, lactic acid, ethanol±water
solution; Lactic acid, ethanol±water solution > saliva, distilled water
Z100 DH Distilled water > citric acid
DL Distilled water, citric acid, lactic acid, heptane, ethanol±water solu-
tion > saliva; Ethanol±water solution > lactic acid
Ariston DH Saliva, distilled water, ethanol-water solution > heptane
pHc DL Saliva > distilled water, citric acid, lactic acid, heptane, ethanol±water
solution; Distilled water, citric acid, lactic acid, heptane > ethanol±water
solution
Sure®l DH NS
DL Saliva, distilled water, citric acid, lactic acid > heptane, ethanol±water
solution; Citric acid, lactic acid > saliva, distilled water

Results on one-way ANOVA and Scheffe's test (P < 0.05); > indicates statistical signi®cance
and NS indicates no signi®cance.

respectively. A negative DH was also observed for
Sure®l specimens which were conditioned in arti®cial
saliva, distilled water and citric acid. Ranking of DH and
DL by media varied for the different composites.
Ranking of DH values by media from most negative
to positive were as follows: Silux ± C < S < L < W <
H < E; Z100 ± C < S < H < L < E < W; Ariston ±
H < C < L < W < S < E; and Sure®l ± C < S < W <
H < E < L. Ranking of DL in lm by media from the
smallest to the largest were as follows: Silux ±
W < S < C < L < E < H; Z100 ± S < L < H < C <
W < E; Ariston ± E < C < H < W < L < S; and Sure®l
± H < E < S < W < C < L.
Signi®cant differences in DH values between differ-
ent media for individual composites are re¯ected in
Table 3. For Silux, conditioning in ethanol±water solu-
tion resulted in signi®cantly greater hardness change
compared with all other chemical media. A signi®cant
difference in DH was also observed between water and
citric acid for Z100 and between heptane and saliva,
water, ethanol±water solution for Ariston. No signi®-
cant difference in DH between media was observed for
Sure®l. For Silux, thickness of DL after conditioning in
heptane was signi®cantly greater than that observed
after conditioning in all other chemical media. In
addition, conditioning in lactic acid and ethanol±water
solution resulted in signi®cantly larger DL than saliva
and water. For Z100, conditioning in arti®cial saliva
resulted in a signi®cantly smaller DL compared with all
other media. A signi®cantly larger DL was observed
after conditioning of Z100 in ethanol±water solution
when compared with lactic acid. Conditioning of

ã 2001 Blackwell Science Ltd, Journal of Oral Rehabilitation 28; 1015±1021

 CHEMICAL DEGRADATION OF COMPOSITE RESTORATIVES 1019

Table 4. Results of statistical analysis based on chemical media Two-way ANOVA of DH and DL showed signi®cant
interaction between materials and chemical media. The
Media Property Difference effects of chemical media on hardness and degradation
Arti®cial saliva DH Ariston > Silux layer thickness are therefore material dependent. A
DL Ariston > Silux, Z100, Sure®l signi®cant but weak positive correlation (correlation
Distilled water DH Z100 > Silux, Sure®l coef®cient r ˆ 0á20) exists between DH and DL.
DL Ariston > Silux, Z100, Sure®l
Citric acid DH NS
DL Ariston > Silux, Z100, Sure®l Discussion
Lactic acid DH Sure®l > Silux
DL Ariston > Silux, Z100, Sure®l The chemical agents used in this study were amongst
Silux, Sure®l > Z100 those recommended in the `FDA Guidelines' (FDA,
Heptane DH NS 1976, 1988) to be used as food simulators. Heptane
DL Silux > Z100, Ariston, Sure®l
simulates butter, fatty meats and vegetable oils. Ethanol
Ariston > Z100, Sure®l
Ethanol±water DH NS solution, citric and lactic acids simulate certain bever-
solution DL Silux > Z100, Ariston, Sure®l ages (including alcoholic ones), vegetables, fruits, can-
Z100 > Ariston, Sure®l dy and syrup. Conditioning in arti®cial saliva was used
Ariston > Sure®l as control. To minimize the oxygen inhibition layer,
Results on one-way ANOVA and Scheffe's test (P < 0.05); which may in¯uence the results, the composites were
> indicates statistical signi®cance and NS indicates no signi®cance. cured against an acetate strip (Finger & Jùrgensen,
1976). The depth of inhibition when composites are
Ariston in arti®cial saliva resulted in signi®cantly exposed to atmospheric air varies between composites
greater DL than in all other mediums. The DL after and ranged from 25 to 105 mlm (Finger & Jùrgensen,
conditioning in heptane, water, citric and lactic acids 1976). Hardness is de®ned as the resistance to perma-
were in turn signi®cantly larger than that after condi- nent indentation or penetration. It is, however, dif®cult
tioning in ethanol±water solution. For Sure®l, condi- to formulate a de®nition that is completely acceptable,
tioning in saliva, water, citric and lactic acids resulted in as any test method will involve a complex interaction of
a signi®cantly greater DL than in heptane and ethanol± stresses in the test material from the force applied.
water solution. DL with citric and lactic acids was Despite this condition, the most common concept of
signi®cantly larger than that with saliva and water. hard and soft substances is the relative resistance to
No signi®cant difference in DH was observed penetration when indented by a hard asperity (McCabe,
between composites when conditioned in citric acid, 1994). Amongst the properties related to hardness of a
heptane and ethanol±water solution (Fig. 1). In arti®- material are proportional limit, ductility, strength and
cial saliva, all composites demonstrated an increase in wear (Anusavice, 1996).
surface hardness with the exception of Ariston. After Chemical softening of restoratives may result in
conditioning in lactic acid, all composites experienced a decreased physico-mechanical properties. The latter
decrease in hardness with the exception of Silux. Both includes diametral tensile strength (Lee et al., 1994)
Ariston and Sure®l had signi®cantly greater DH than and fracture toughness (Ferracane & Marker, 1992),
Silux. Z100 underwent a signi®cantly greater change in hardness and wear (Wu & McKinney, 1982). As the
hardness compared with Silux and Sure®l after condi- greatest change in hardness had been shown to occur
tioning in water. The only medium, which resulted in a within the ®rst 7 days (Kao, 1989) and hardness of
consistent decrease in hardness of all composites after composites is affected by conditioning for 7 days
conditioning, was 75±25% ethanol±water solution. (McKinney & Wu, 1985), this period of conditioning
After conditioning in arti®cial saliva, water, citric and was selected for this experiment. As the restorative
lactic acids, Ariston had signi®cantly larger DL than the materials were not exposed to any mechanical forces,
other composites (Fig. 2). The largest DL (43á95 lm) any observed changes in hardness would be from
was observed after conditioning Ariston in arti®cial chemical reaction or dissolution. The 24-h storage
saliva. After conditioning in organic solvents (i.e. period in arti®cial saliva before hardness testing was
heptane and ethanol±water solution), Silux had signi- essential for elution of unreacted components from the
®cantly larger DL than all other composites. composites (Ferracane & Condon, 1990) and to allow

ã 2001 Blackwell Science Ltd, Journal of Oral Rehabilitation 28; 1015±1021

1020 A . U . J . Y A P et al.
for composite post-cure, if any, to occur (Yap, 1997). A
negative DH, which indicates an increase in hardness,
was observed with Silux after conditioning in all media
with the exception of ethanol±water solution. A neg-
ative DH was also observed for Z100 conditioned in
arti®cial saliva and citric acid, Ariston conditioned in
heptane, and Sure®l conditioned in arti®cial saliva,
water and citric acid. This increase in hardness over the
1 week conditioning period re¯ects a progressive cross-
linking reaction that occurs following light-activation of
composite resins (Watts et al., 1987). This suggests
appreciable residual concentration of free radicals in
the composites after light polymerization.
When Silux specimens were stored in 75±25%
ethanol±water solution, signi®cant softening was noted
when compared with all other chemical media. This
solution has been the solvent of choice to simulate
accelerated ageing of restorations as it has a solubility
parameter, which matches that of BisGMA (Wu &
McKinney, 1982; McKinney & Wu, 1985). In the
present study, ethanol±water solution was the only
media that consistently softened all four BisGMA-based
composite resins (Fig. 1). The thickest degradation layer
was, however, observed after conditioning Silux in
heptane which being an organic solvent, like ethanol±
water solution, has the potential for polymer damage.
The thick degradation layer did not appear to affect
hardness. Silux had signi®cantly greater DL than all
other composites after conditioning in both organic
chemical liquids, i.e. heptane and ethanol±water solu-
tion. This can be explained in part by the higher resin
content of Silux compared with the other composites.
Z100 was the most susceptible to the softening effect
of water and had the largest DH for all material±media
combination (Fig. 1). A signi®cant softening of Z100
was observed after conditioning in water as compared
with citric acid. Surface hardness of composites was
found to be signi®cantly affected by both water sorption
and the contact time with the aqueous media (Hansen,
1983). In addition to the effects upon the resin matrix,
degradation of the resin±®ller interface and inorganic
®llers may also play a role in the reduction of surface
hardness (SoÈderholm, 1981). The latter is the more
favoured explanation for the result of Z100 as all four
composites investigated were BisGMA-based. Z100
utilizes synthetic zirconia/silica ®llers, which like other
silica ®llers have irregularly distributed Si±O±Si bonds.
Swelling from water sorption by the resin matrix could
induce stress around the stiff ®ller inclusions as a result
ã 2001 Blackwell Science Ltd, Journal of Oral Rehabilitation 28; 1015±1021
of matrix expansion. The high energy levels resulting
from strained Si±O±Si bonds, makes the ®llers more
susceptible to stress corrosion attack resulting in com-
plete or partial debonding of ®llers (SoÈderholm, 1983)
which decreases hardness. However, not all glasses are
affected to the same degree. Composites containing zinc
and barium glasses have been shown to be more
susceptible to aqueous attack compared with quartz
(SoÈderholm, 1983; Oysñd & Ruyter, 1986). From the
present experiment, it appears that zirconia glasses are
also susceptible to aqueous attack. This may be com-
pounded by the smaller ®ller surface area associated
with the spherical shape of zirconia/silica ®llers that
may decrease bonding of ®llers to the resin matrix. The
DH of Z100 was signi®cantly greater than that of Silux
and Sure®l after conditioning in water. The degradation
layer of Z100 after conditioning in saliva was signi®-
cantly lower than that in all other media and condi-
tioning in ethanol±water solution resulted in larger DL
values than conditioning in lactic acid. DL range was,
however, small and varied from 3á12 in arti®cial saliva
to 7á55 for ethanol±water solution.
For Ariston, conditioning in heptane resulted in
signi®cantly different DH values compared with saliva,
water and ethanol±water solution. Conditioning in
heptane resulted in an increase in hardness. Two
possible explanations are that heptane reduces oxygen
inhibition during post-curing and eliminates leaching
out of silica and combined metal in ®llers, which occurs
from conditioning in aqueous solutions (SoÈderholm,
1983). The signi®cant positive effect of heptane with
regard to hardness change was, however, noted only
with Ariston. This may be attributed to the open
structure of Ariston that is required for more effective
¯uoride and ionic release (Combe & Douglas, 1998).
This open structure increases solvent diffusion (which is
reduced by heptane) that also explains the signi®cantly
larger degradation layer noted with Ariston as com-
pared with all the other composites after conditioning
in arti®cial saliva (which consists mainly of water),
water, citric and lactic acids (Fig. 2). In addition, the
degradation layer of Ariston in ethanol±water solution
and heptane was signi®cantly larger than Sure®l, and
Z100 and Sure®l, respectively. The degradation layer of
Ariston in arti®cial saliva was signi®cantly greater than
that in all other media.
For Sure®l, no signi®cant difference in DH was
observed after conditioning in the different chemical
media. After conditioning in lactic acid, Sure®l had a
 CHEMICAL DEGRADATION OF COMPOSITE RESTORATIVES
signi®cantly larger DH than Silux. Lactic acid may cause
chemical dissolution of the ¯uorosilicate glass ®llers in
Sure®l, as weak intra-oral acids have been shown to
cause vast chemical dissolution of glass±ionomer
cements (Roulet & Walti, 1984). The degradation layer
of Sure®l after conditioning in arti®cial saliva, water,
citric and lactic acids was signi®cantly greater than that
of organic solvents. Amongst the fore mentioned group,
DL values obtained with citric and lactic acids was
signi®cantly greater than saliva and distilled water.
Two-way ANOVA of DH and DL showed that the
effects of chemical medium on hardness and degrada-
tion layer thickness is material dependent. The ranking
of DH and DL by chemical media were all different for
composites evaluated. The positive correlation between
DH and DL was signi®cant but very weak as correlation
coef®cient r was only 0á20. The changes in surface
hardness of composite resins are therefore weakly
associated with the thickness of the degradation layer.
A thicker degradation layer is coupled with greater
softening of the composite surface.
Conclusions
Within the limitations of this study, we may conclude
that:
1. Both organic solvents and water/weak acids may
degrade composite resins.
2. The effects of chemical media on change in surface
hardness and thickness of degradation layer is material
dependent.
3. All composites are softened by 75±25% ethanol±
water solution.
4. A thicker degradation layer is weakly associated with
greater softening of composites.
References
ANUSAVICE , K.J. (1996) Mechanical properties of materials. In:
Skinner's Science of Dental Materials, 10th edn (ed. K.J.
Anusavice), p. 69. W.B. Saunders Company, Philadelphia,
USA.
COMBE , E.C. & DOUGLAS , W.H. (1998) The future of dental
materials. Dental Update, 25, 411.
FERRACANE , J.L. & CONDON , J.R. (1990) Rate of elution of
leachable components from composite. Dental Materials,
6, 282.
FERRACANE , J.L. & MARKER , V.A. (1992) Solvent degradation and
reduced fracture toughness in aged composites. Journal of Dental
Research, 71, 13.
ã 2001 Blackwell Science Ltd, Journal of Oral Rehabilitation 28; 1015±1021
1021
FINGER , W. & JùRGENSEN , D.S. (1976) Inhibition of polymerization
by oxygen in composite ®lling materials and enamel sealers.
Schweizerische Monatsschrift Fur Zahnheilkunde, 86, 812.
FOOD AND DRUG ADMINISTRATION, BUREAU OF FOODS
(1976) Guidelines for Chemistry and Technology, p. 11. Department
of Health, Education and Welfare, Washington DC, USA.
FOOD AND DRUG ADMINISTRATION, DIVISION OF FOOD
CHEMISTRY AND TECHNOLOGY (1988) Recommendations for
Chemistry Data for Indirect Food Additives Petitions, p. 9. Depart-
ment of Health and Human Services, Washington DC, USA.
HANSEN , E.K. (1983) After polymerization of visible light activated
resins: surface hardness vs. light sources. Scandinavian Journal of
Dental Research, 93, 280.
KAO , E.C. (1989) In¯uence of food-simulating solvents on resin
composites and glass-ionomer restorative cement. Dental Mate-
rials, 5, 201.
LEE , S.Y., GREENER , E.H., MUELLER , H.J. & CHIU , C.H. (1994)
Effect of food and oral simulating ¯uids on dentin bond and
composite strength. Journal of Dentistry, 22, 175.
LEE , S.Y., GREENER , E.H. & MENIS , D.L. (1995) Detection of
leached moieties from dental composites in ¯uids simulating
food and saliva. Dental Materials, 11, 348.
MC CABE , J.F. (1994) Properties used to characterize materials. In:
Applied Dental Materials, 7th edn (ed. J.F. Mccabe), p. 12.
Blackwell Scienti®c Publications, Oxford, UK.
MC KINNEY , J.E. & WU , W. (1985) Chemical softening and wear of
dental composites. Journal of Dental Research, 64, 1326.
OYSáD , H. & RUYTER , I.E. (1986) Water sorption and ®ller
characteristics of composites for use in posterior teeth. Journal of
Dental Research, 65, 1315.
PHILLIPS , R.W. (1981) Past, present and future of composite
systems. Dental Clinics of North America, 25, 209.
ROULET , J.F. & WALTI , C. (1984) In¯uence of oral ¯uid on
composite resins and glass ionomer cement. Journal of Prosthetic
Dentistry, 52, 182.
SoÈDERHOLM , K.J. (1981) Degradation of ®llers in experimental
composites. Journal of Dental Research, 60, 1867.
SoÈDERHOLM , K.J. (1983) Leaking of ®llers in dental composites.
Journal of Dental Research, 62, 126.
VAN GROENINGEN , G., JONGEBLOED , W. & ARENDS , J. (1986)
Composite degradation in vivo. Dental Materials, 2, 225.
WATTS , D.C., AMER , O., & COMBE , R.C. (1987) Surface hardness
development in light-cured composites. Dental Materials, 3, 265.
WU , W. & MC KINNEY , J.E. (1982) In¯uences of chemicals on wear
of dental composites. Journal of Dental Research, 61, 1180.
WU , W., TOTH , E.E., MOFFA , J.F. & ELLISON , J.A. (1984) Subsur-
face damage layer of in vivo worn dental composite restorations.
Journal of Dental Research, 63, 675.
YAP , A.U.J. (1997) Post-irradation hardness of resin-modi®ed glass
ionomer cements and a polyacid-modi®ed composite resin.
Journal of Material Science, 8, 413.
Correspondence: Dr Adrian U. J. Yap, Department of Restorative
Dentistry, Faculty of Dentistry, National University of Singapore, 5
Lower Kent Ridge Road, Singapore 119074.
E-mail: rsdyapuj@nus.edu.sg