Dent Mater8:362-365, November,1992
Evaluation of new methods for composite repair
E.J. Swift, Jr. 1, B.D. LeValley 2, D.B. Boyer 1
1Department of Operative Dentistry, The University of Iowa, Iowa City, IA, USA
2General practice resident, St. Joseph Hospital, Denver, CO, USA
Abstract. This study evaluated the effectiveness of air abrasion
(sandblasting), hydrofluoric acid (HF) etching, and acidulated
phosphate fluoride (APF) for composite repair. Five different
composites (three hybrids, one small-particle, and one microfill)
were used to test the effects of different compositions and particle
sizes. A phosphonate resin bonding agent was used for all repair
procedures. In general, air abrasion provided the strongest
repairs, approximately 60% of cohesive strength. The effect of HF
varied with different composites, either improving or reducing the
repair strength. APF gave the lowest interfacial bond strengths for
all composites except the microfill.
INTRODUCTION
Complete removal of defective composite restorations is not
always necessary or desirable. Repairs of fractured restora-
tions, resurfacing of discolored restorations, and partial re-
placement of deep or complex restorations are frequently
accomplished by the addition of new composite to the existing
material. Also, composite restorations are sometimes ve-
neered with direct composite or resin-bonded etched porce-
lain.
The interfacial bonding strength of composite to fresh (i.e.,
only a few minutes old) composite is the same as the cohesive
strength of the material (Lloyd et al., 1980; Boyer et al., 1984).
However, repairs to ground, untreated composite surfaces
usually have bond strengths in the range of 20% to 70% of
cohesive strength (Lloyd et al., 1980; Vankerckhoven et al.,
1982; Chan and Boyer, 1983; Boyer et al., 1984; Azarbal et al.,
1986; Kao et al., 1988; Puckett et al., 1991).
The number of unreacted methacrylate groups decreases
as the resin polymerizes, so the potential for primary bonding
to these groups diminishes over time. Furthermore, when the
composite surface is ground or cut, inorganic filler particles (in
hybrid and small-particle composites) or pre-polymerized
particles (in microfills) are exposed. Primary resin bonding to
these particles is not likely (Boyer et al., 1984).
A variety of surface treatments and bonding agents has
been used to improve the repair strength of composites.
Unfilled resins improve bonding of fresh material to ground
composite surfaces (Boyer et al., 1984; Azarbal et al., 1986;
Puckett et al., 1991). Phosphonate ester dentin/enamel bond-
ing agents enhance the bond to a greater extent (Azarbal et al.,
1986; Kao et al., 1988; Puckett et al., 1991). Dissimilar
composites (e.g., Bis-GMA or urethane dimethacrylate) gen-
erally bond as well as similar composites (Kao et al., 1988;
Gregory et al., 1990), so repair strength is not improved by
using materials with the same resin formulation.
362 Swift et aL/Composite repair
Hydrofluoric acid (HF) is commonly used to etch porcelain
surfaces for indirect restorations or intraoral repairs. Acidu-
lated phosphate fluoride (APF) can be used, but etches less
effectively(Lacy et al., 1988; A1 Edris et al., 1990). Both acids
also etch the glass filler particles in hybrid and small particle
composites. Although microfill composites do not contain
glass particles, HF and APF can cause some slight surface
changes (Kula et al., 1983; Kula et al., 1986). Some clinicians
believe that etching of glass filler particles creates a mechani-
cally retentive surface, improving the repair strength of
composites. Etching times of 2-10 min for APF and 30 s orless
for HF have been recommended (Miller, 1990; Hamilton,
1990).
Research regarding the effectiveness of HF and APF for
composite repair is rather limited. Crumpler et al. (1989)
reported that neither agent significantly improved the repair
strength of posterior composites. Mitchem et al. (1991) re-
cently recommended that hybrid composites should not be
etched because etching causes softening and porosity in the
composite surface.
The purpose of our study was to evaluate the effectiveness
of composite repair using etching gels and a mechanical
roughening technique (air abrasion) under controlled in vitro
conditions. Five composites were used to test the effects of
different filler compositions and particle sizes.
MATERIALS AND METHODS
The composites used in this study are described in Table 1.
Rectangular composite specimens were made in a split alumi-
num mold 2.0 mm deep, 2.5 mm wide, and 15 mm long. The
mold was lightly lubricated with petroleum jelly. A Teflon
spacer was inserted to make half-length specimens. The
composite material was packed into the mold, covered with a
Mylar strip and glass microscope slide, and compressed with
a 1 kg weight. A wide-tip visible light-curing unit (Prismetics,
Caulk/Dentsply, Milford, DE, USA) was used to polymerize
the resin. An exposure time of 40 s was used for half-length
specimens. Two overlapping 40 s exposures were used for full-
length specimens. Thirty half-length specimens and ten full-
length specimens of each composite were made. The full-
length specimens were used to determine the fracture strength
of intact, unrepaired composites.
The half-length specimens were mounted in a jig, and
polished sequentially with 320, 400, and 600 grit silicon
carbide paper. All specimens were stored in distilled water for
7 d at room temperature after finishing. After storage, 10
specimens of each composite were treated with the following
 TABLE 1: COMPOSITES USED IN THIS STUDY specimen was calculated using the fol-
lowing formula (Craig, 1989):
Composite Manufacturer Classification Primary Filler Mean Particle
Size
TS = :/d2 P
Herculite XRV Kerr Manufacturing hybrid barium glass 0.6 pm
Romulus, MI, USA colloidal silica 0.04 IJm where P = fracture load (kg), l = sample
length between supports (10 mm), b =
P-50 3M Dental Products hybrid zirconia/silica 1.5 pm sample width (2.5 ram), and d = sample
St. Paul, MN, USA colloidal silica 0.04 IJm thickness (2 mm).
The data were analyzed using analy-
Pertac-Hybrid ESPE-Premier hybrid quartz 1.2 pm sis ofvariance and Duncan multiple range
Norristown, PA, USA colloidal silica 0.04 pm tests in the SAS statistical software
yttrium fluoride opacifier package (SAS Institute, Cary, NC,
USA). Additional specimens of compos-
Prisma-Fil Caulk/Dentsply small particle barium glass 4.0 IJm ite were fabricated, treated, and exam-
Milford, DE, USA ined using scanning electron microscopy
(SEM) to evaluate changes in surface
Silux Plus 3M Dental Products microfill colloidal silica 0.04 pm topography.
RESULTS
The results ofthis study are summarized
TABLE 2: SUMMARY OF TRANSVERSE BOND STRENGTHS (MPA) OF REPAIRED COMPOSITES*
in Table 2. A two-way analysis of vari-
Mean (S.D.) ance (ANOVA) revealed significant dif-
Surface Treatment Pooled HerculiteXRV P-50 Pertac-Hybrid Prisma-Fil Silux Plus ferences related to material, surface treat-
ment, and the interaction of these two
1. Control (intact) 92.6 (12.9) 99.8 (10.1) 85.1 (18.8) 111.6 (15.0) 105.2 (13.7) 61.4 (6.7) variables (all at p < 0.0001). Multiple
one-way ANOVA were performed to
2. Air Abrasion 55.0 (24.9) i 68.2 (22.9) I 38.5 (18.3) 67.3 (25.0) 59.3(34.5) 41.5 (12.0) evaluate the repair methods overall and
for each material. Duncan's multiple
3.HFEtch 50.9(18.6) 71.1 (19.8) 45.2(14.1) 41.9 (16.7) 50.3 (29.0) 45.7 (13.2)!
range test (p = 0.05) was used for pair-
4.APF E t c h 39.3(14.7) 62.6(21.4) 28.8 (9.5) 24.6(20.7)[ 38.2 (9.1) 44.4 (12.9) i wise comparisons where indicated.
The cohesive strength of intact speci-
*Values connected by vertical lines are not significantly different at p < 0.05. mens ranged from 61.4 MPa for the
microfill composite, Silux Plus, to
methods:Airabrasion: The experimental surface was briefly 111.6 MPa for Pertac-Hybrid, a quartz-filled composite. The
(5-10 s) air-abraded with 50 ~m aluminum oxide particles highest absolute repair strength was 71.1 MPa, for air-abraded,
using an intraoral sandblasting device (Microetcher, Danville HF-etched Herculite XRV. This value represents 71% of the
Engineering, Danville, CA, USA), then rinsed and dried. HF cohesive strength of Herculite.
gel: The polished surface was briefly air-abraded; 9.6% The highest relative repair strength was 45.7 MPa for
hydrofluoric acid gel (Porcelain Etch Gel, Pulpdent Corp., sandblasted, HF-etched Silux Plus, which was 74% of that
Watertown, MA, USA) was applied to the ground surface for material's cohesive strength. When the data for all composites
30 s, then rinsed with water for 10-15 s and dried with were pooled, the mean strengths of the different repair meth-
compressed air. APFgel: After brief air abrasion, 1.23% APF ods were in the range of 40-60% of cohesive strength, and all
gel (Nupro, Johnson & Johnson Dental Care Co., New were significantly less than the mean cohesive strengths. Air
Brunswick, NJ, USA) was applied for 10 rain and removed in abrasion generally provided the best repair strengths (55.0 _+
the same manner. 24.9 MPa). HF etching gave slightly lower, but not signifi-
Each half-length specimen was placed back into the mold. cantly different, interfacial bond strengths (50.9 _+18.6 MPa).
A phosphonate ester dentin/enamel bonding agent (Bondlite, APF provided significantly weaker repairs than either of the
Kerr Manufacturing Company, Romulus, MI, USA) was ap- other two methods (39.3 _ 14.7 MPa).
plied, thinned with compressed air, and cured with a 20 s For three of the composites (Herculite XRV, Prisma-Fil,
exposure to visible light. Fresh composite was placed and and Silux Plus), the various repair methods were statistically
polymerized in the mold to complete the repair procedure. The similar. For P-50, HF etching provided the strongest and APF
repair and substrate materials were the same; e.g., Herculite etching provided the weakest repairs. The highest bond
XRV was repaired with Herculite XRV. The repaired speci- strengths for Pertac-Hybrid were obtained with air abrasion.
mens were stored in distilled water for 24 h before bond Both etching techniques, particularly APF, gave very low
strength testing. repair strengths for this material.
The transverse strength of all specimens was tested with Evaluation oftreated composite surfaces with SEM showed
a three-point apparatus in a universal testingmachine (Instron that air abrasion caused substantial roughening of all four
model TT-D, Instron Corp., Canton, MA, USA). The length materials (Fig. 1). The effects ofHF and APF were less obvious,
between supports was 10 ram, with the joint placed directly although both agents apparently removed portions ofthe resin
under the center edge of the loading probe. Cross-head speed matrix as well as some filler particles from the small-particle
was 0.5 cm/min. The transverse strength (MPa) of each and hybrid composites (Fig. 2).

Dental Materials~November 1992 363

Fig. 1. Surfacesof composites which wereair-abraded for lOswith 50 pmaluminum oxide: (a) microfilI -- Silux Plus; (b)glass-filled hybrid-- Herculite XRV;
(c) quartz-fUled hybrid -- Pertac-Hybrid; and (d) hybrid posterior composite -- P-50.
Fig. 2. (a) High magnification (lO,O00x) view of air-abraded Herculite XRV barium glass-filled hybrid composite. (b) Hercuiite XRV which was air-abraded, then etched for
30 s with HF.
DISCUSSION
Of the composites tested, Herculite XRV had the highest
repair strengths. However, when interfacial bond strengths
were expressed as percentages of cohesive strength, the
microfill composite, Silux Plus, had the most effective repairs.
Silux Plus has the highest resin content of the materials
tested, and therefore is more easily abraded and has more
potential sites for resin-resin bonding than the heavily-filled
materials. Also, Silux Plus has a relatively low viscosity, and
may wet the substrate better than more highly-filled and
viscous composites.
In general, air abrasion with 50 ~m aluminum oxide
particles provided the highest composite repair strengths.
Repairs using air abrasion had interfacial bond strengths
ranging from 45% to 70% of cohesive strength. These values
are in the typical range reported for composite repairs (Boyer
et al., 1984; Kao et al., 1988; Crumpler et al., 1989). However,
in comparing our results with those of Crumpler et al. (1989),
who studied other methods of mechanical roughening, air
364 Swift et al./Composite repair
abrasion apparently is no more effectivethan roughening with
a diamond or carbide bur.
Etching with a 9.6% HF gel for 30 s slightly increased the
repair strength of three composites (Herculite XRV, P-50, and
Silux Plus). However, the increased strength was not signifi-
cant for any of the composites. Herculite XRV and P-50 both
contain glass filler particles which could be etched by HF.
Silux Plus is a microfill material, and therefore does not
contain etchable fillers. HF decreased the repair strengths of
Prisma-Fil and Pertac-Hybrid. The decrease for Pertac-
Hybrid was particularly great and also statistically signifi-
cant. Pertac-Hybrid is heavily filled with non-etchable quartz
rather than glass particles, as in the other hybrids tested.
Also, the resin matrix of this material contains a more hydro-
phobic monomer than that of most composites.
HF etching has been advocated as a method for in-
creasing composite repair strengths. However, our results
indicate that HF etching does not significantly improve re-
pairs. In fact, it may adversely affect repairs of some compos-
ites. The failure of HF to consistently increase bonding of
composite to itself or to tooth structure has been shown in
other studies. HF apparently causes porosity and softening of
the composite surface (Crumpler et al., 1989; Mitchem et al.,
1991; Swift et al., 1992). Given that the composition of
substrates requiring repair is frequently unknown, clinicians
may wish to avoid routine use of HF gels in composite repair
procedures.
Etching with a 1.23% APF gel provided the lowest bond
strengths overall and for most of the specific composites. APF
gels have a high viscosity, perhaps making complete removal
difficult. These gels also include thickeners, flavoring agents,
and other contents which would have an unknown effect on
composite substrates. In addition, the 10-minute application
time used in our study might be excessive. Nevertheless, our
results indicate that APF gels should not be used for composite
repair.
Silane coupling agents improve the bond of composite to
abraded or etched porcelain (Stangel et al., 1987; Lacy et al.,
1988) and might improve repairs of air-abraded or etched
composite. However, this hypothesis was not tested in this
study.
In summary, air abrasion with 50 ~m aluminum oxide
particles provided a mechanically retentive surface for com-
posite repairs. Etching composite with a 9.6% HF gel for 30 s
either increased or decreased repair strength, depending on
the composite material being repaired. Etching composite
with 1.23% APF gel for 10 min reduced the repair strength
obtained by air abrasion alone. There was no correlation
between filler particle size and repair strength of the compos-
ite materials tested.
ACKNOWLEDGMENTS
We thank Dr. Jorge Perdigao for his assistance with the
scanning electron microscopy.
Materials for this project were provided by Kerr Manufac-
turing Company (Romulus, MI, USA), 3M Dental Products
Division (St. Paul, MN, USA), ESPE-Premier Sales Corpora-
tion (Norristown, PA, USA), Caulk/Dentsply (Milford, DE,
USA), and Pulpdent Corporation (Watertown, MA, USA).
Received July 1, 1992/AcceptedOctober 18, 1992
Address correspondence and reprint requests to:
Dr. Edward J. Swift, Jr.
Department of Operative Dentistry
The University of Iowa
Iowa City, IA 52242 USA
REFERENCES
AI Edris A, Al Jabr A, Cooley RL, Barghi N (1990). SEM
evaluation of etch patterns by three etchants on three
porcelains. J Prosthet Dent 64:734-739.
Azarbal P, Boyer DB, Chan KC (1986). The effect of bonding
agents on the interfacial bond strength ofrepaired compos-
ites. Dent Mater 2:153-155.
Boyer DB, Chan KC, Reinhardt JW (1984). Build-up and
repair oflight-cured composites: Bond strength. JDent Res
63:1241-1244.
Chan KC, Boyer DB (1983). Repair of conventional and
microfilled composite resins. J Prosthet Dent 50:345-350.
Craig RG (1989). Restorative Dental Materials. 8th ed. St.
Louis: Mosby, 88.
Crumpler DC, Bayne SC, Sockwell S, Brunson D, Roberson
TM (1989). Bonding to resurfaced posterior composites.
Dent Mater 5:417-424.
Gregory WA, Pounder B, Bakus E (1990). Bond strengths of
chemically dissimilar repaired compositeresins. JProsthet
Dent 64:664-668.
Hamilton JC (1990). Kerr products questions and answers.
2nd ed. Romulus, MI: Kerr Manufacturing Company, 54.
Kao EC, Pryor HG, Johnston WM (1988). Strength of compos-
ites repaired by laminating with dissimilar composites. J
Prosthet Dent 60:328-333.
Kula K, Nelson S, Thompson V (1983). In vitro effect of APF
gel on three composite resins. JDent Res 62:846-849.
Kula K, Nelson S, Kula T, Thompson V (1986). In vitro effect
of acidulated phosphate fluoride gel on the surface of
composites with different filler particles. J Prosthet Dent
56:161-169.
Lacy AM, LaLuz J, Watanabe LG, Dellinges M (1988). Effect
of porcelain surface treatment on the bond to composite. J
Prosthet Dent 60:288-291.
Lloyd CH, Baigrie DA, Jeffrey IW (1980). The tensile strength
of composite repairs. JDent 8:171-177.
Miller, MB, ed. (1990) Repairing composites. Reality 5:216.
Mitchem JC, Ferracane JL, Gronas DG (1991). The etching
of hybrid composite to facilitate cementation or repair. J
Dent Res 70:392, Abstr. No. 1007.
Puckett AD, Holder R, O'Hara JW (1991). Strength of poste-
rior composite repairs using different composite/bonding
agent combinations. OperDent 16:136-140.
Stangel I, Nathanson D, Hsu CS (1987). Shear bond strength
of the composite to etched porcelain. JDent Res 66:1460-
1465.
Swift EJ, Brodeur C, Cvitko E, Pires JAF (1992). Treatment
of composite surfaces for indirect bonding. Dent Mater
8:193-196.
Vankerckhoven H, Lambrechts P, Van Beylen M, Davidson
CL, Vanherle G (1982). Unreacted methacrylate groups on
the surface of composite resins. JDent Res 61:791-795.
Dental Materials~November 1992 365