Flexural fatigue behavior of resin composite dental restoratives

Ulrich Lohbauera,*, Tina von der Horstb, Roland Frankenbergera,Norbert Kramera, Anselm Petschelta

aPoliclinic for Operative Dentistry and Periodontology, University of Erlangen-Nuremberg, Gluckstrasse 11, 91054 Erlangen, GermanybDepartment of Medical Informatics, Biometrics, and Epidemiology, University of Erlangen-Nuremberg, Erlangen, Germany

Received 22 February 2002; revised 13 June 2002; accepted 31 July 2002

Abstract

Objectives. The aim of this study was to evaluate the mechanical properties of resin composite dental restoratives under quasi-static and

cyclic loading.

Methods. Four-point-bending bars of 10 different resin composite materials were manufactured according to ISO standard and stored for

two weeks in distilled water. The fracture strength (FS) was measured with the four-point-bending test in an universal testing machine. The

flexural fatigue limits (FFL) for 105 cycles were determined under equivalent loading. All specimens were tested and fatigued in water at

37 8C. The data were analyzed using ANOVA, Weibull statistics of FS and the ‘staircase’ approach of FFL. Fractographic analysis was

performed using SEM.

Results. The initial flexural strength values for the resin composite materials varied from 55.4 MPa for Solitairew up to 105.2 MPa for

Filtekw Z250. The mean flexural fatigue limit for 105 cycles ranged between 37 and 67% of the initial strength. SEM analysis of the fractured

surfaces suggests two kinds of failure mechanisms for initial and fatigue fracture.

Significance. The fatigue behavior of resin composite materials does not correlate with initial strength values. Materials providing high

initial strengths do not obviously reveal the best fatigue resistance. Flexural fatigue measurement of resin composite materials should be

viewed as a useful tool to evaluate long term mechanical properties.

q 2003 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved.

Keywords: Fracture strength; Fatigue; Resin composite; In vitro; Staircase; Mechanical properties

1. Introduction

Within the last few decades, modern restorative materials

were developed with a focus on amalgam-like mechanical

properties, excellent aesthetics and biocompatibility. Such

materials were further improved for application in stress

bearing areas. Therefore, mechanical properties under

masticatory load and above all fatigue resistance are

important.

Fatigue fractures after years in clinical use were found to

be a common failure reason. Damage of restorations like

bulk, cusp, or marginal fractures were frequently reported

[1,2]. Using resin composite materials, Burke et al. [3]

reported marginal fracture (18%) and bulk fracture (7%) as

the most prevalent reasons for rerestoration.

Fatigue in dental restoratives is influenced by corrosive

water attack at a certain temperature (37 8C) and by cyclic

masticatory forces. The naturally occuring loading of a

filling was estimated at between 5 and 20 MPa [4].

Contemporary approaches to fatigue principles consider a

fracture process in three phases: crack initiation, slow crack

growth, and fast fracture. The latter phase is very short in

duration and thus the time of crack initiation and of slow

crack growth account for the useful fatigue resistance of a

material. Crack initiation nucleates at heterogenities like

surface and subsurface microcracks, porosities, filler

particles, crazes, etc. within the material [5]. Cyclic loading

is able to drive a crack, called slow crack growth. Additional

water exposure causes a variety of weakening effects on

resin composites: degradation of the filler–matrix interface,

eluation, and swelling or a visco-elastic effect on the matrix

which all accelerate slow crack growth [6–8].

The purpose of this in vitro study was to determine the strength

of today’s resin composite materials under fatigue conditions,

0109-5641/03/$ - see front matter q 2003 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved.

doi:10.1016/S0109-5641(02)00088-X

Dental Materials 19 (2003) 435–440

www.elsevier.com/locate/dental

* Corresponding author. Tel.: þ 49-9131-853-4236; fax: þ49-9131-853-

3603.

E-mail address: lohbauer@dent.uni-erlangen.de (U. Lohbauer).

simulating the clinical situation. The methodology was developed

considering earlier findings on material degradation.

2. Materials and methods

2.1. Materials

A variety of present commercially available, light-curing

resin composite materials were used in this study (Table 1).

They were chosen to be representative for the differences in

filler configuration.

2.2. Specimen preparation

Depending on the material’s density, around 0.2 g was

weighed and placed in a special mold (2 £ 2 £ 25 mm3).

The light polymerization was performed with a halogen

light curing unit (Transluxw CL (800 mW/cm2), Heraeus

Kulzer, Germany) on five overlapping points on each upper

and lower side. The illumination time on a single point was

20 s for Filtekw Z250 and 40 s for the remaining materials.

The procedure followed the manufacturers’ recommen-

dation and ISO 4049 standard. The specimens’ surfaces

were ground with silicon carbide paper up to 800 grit, to

avoid and remove cracks at their edges. All specimens were

stored for two weeks in distilled water at 37 8C.

2.3. Experimental procedure

To evaluate the initial flexural strength the four-point-

bending test was used (n ¼ 12). Bars of 25 mm in length

were fixed between four fins (B ¼ 2 mm, distance of inner

fins ¼ 10 mm, distance of outer fins ¼ 20 mm) and were

subsequently loaded until fracture with a crosshead speed of

0.75 mm/min (universal testing machine Zwicki, Zwick,

Ulm, Germany). The tests were carried out under distilled

water at a temperature of 37 8C.

The flexural fatigue limits (FFL) of the composite

materials were determined for 105 cycles under equivalent

test conditions at a frequency of 0.5 Hz (n ¼ 20). The

‘staircase’ approach method [9,10] was used for fatigue

evaluation. For every cycle the stress alternated between

1 MPa and the maximum stress. Tests were conducted

sequentially, with the maximum applied stress in each

succeeding test being increased or decreased by a fixed

increment, according to whether the previous test resulted in

failure or not. The first specimen was tested at approxi-

mately 50% of the initial flexural strength value. As the data

are concentrated around the mean stress, the number of

specimens required is less than with other methods [11,12].

Fractographic examination was performed under a light

microscope (SV11, Zeiss, Germany) on all specimens and

under a scanning electron microscope (SEM, Leitz ISI SR 50,

Akashi, Japan) on representative specimens (five for each set).

2.4. Statistical treatment

According to the assumption of the weakest link, the

fracture strength (FL) of brittle materials will be limited by

the longest crack size in the loaded volume. Hence, a

distribution of crack lengths results in a strength distribution

which is commonly described by fracture probability P(Fsc)

PFðscÞ¼ 1 2 exp 2

sc

s0

� �m� �ð1Þ

where s0 is the scale parameter (PF(sc) ¼ 63.2%) and m is

the Weibull modulus, respectively [13,14]. The strength

data was evaluated according to this two parameter

cumulative Weibull distribution by plotting the fracture

probability PF(sc)versus fracture strength sc

ln ln1

ð1 2 PFðscÞÞ¼ m ln sc 2 m ln s0 ð2Þ

The parameters m and s0 were determined by a maximum

likelihood approach.

The groups among each other were analyzed using the

non-parametric Mann – Whitney U-Test (a ¼ 0.05/8

¼0.00625; SPSS 10.0 for Windows). Commonly this test

analyses two groups. However, to compare eight groups a

Bonferroni correction was applied.

The mean FFL for 105 cycles was determined using

Eq. (3) and standard deviation, respectively, using Eq. (4)

FFL ¼ X0 þ d

PiniPni

^ 0:5

� �ð3Þ

SD ¼ 1:62d

Pni

Pi2ni 2

Pini

� �2Pni

� �2 þ 0:029

!ð4Þ

Table 1

Resin composite materials under investigation

Brand name (LOT, shade) Manufacturer Composite material Filler fraction (wt%/vol%)

Charismaw (020030, A2) Heraeus Kulzer (Germany) Fineparticle hybrid 78/61

Definitew (218, A2) Degussa (Germany) Fineparticle hybrid, ormocer 78/61

Filtekw Z250 (OCY 2002-10, A3) 3M ESPE (USA) Fineparticle hybrid 82/60

Heliomolarw (905566, A3) Vivadent (Liechtenstein) Inhomogenously microfilled 76.5/64

Solitairew (22, A2) Heraeus Kulzer (Germany) Porous silica hybrid 66/46

Solitairew II (010221, A2) Heraeus Kulzer (Germany) Porous silica hybrid 75/58

Surefilw (9804205, A3) Caulk/Dentsply (USA) Fineparticle hybrid 82/65.2

Tetricw Ceram (A07612, A3) Vivadent (Liechtenstein) Fineparticle hybrid 78.6/60

U. Lohbauer et al. / Dental Materials 19 (2003) 435–440436

where X0 is the lowest stress level considered in the analysis

and d is the fixed stress increment. To determine the FFL,

the analysis of the data is based on the least frequent event

(failures versus non-failures). In Eq. (3) the negative sign is

used when the analysis is based on failures, otherwise the

positive sign is used. The lowest stress level considered is

designated as i ¼ 0, the next as i ¼ 1, and so on and ni is the

number of failures or non-failures at the given stress level.

Analogously to the initial strength comparison, the

fatigue data were analyzed using the non-parametric

Mann–Whitney U-Test (a ¼ 0.00625; SPSS 10.0 for

Windows).

3. Results

The mean flexural strengths s0 at a fracture probability

PFðscÞof 63.2%, the Weibull modulus m and the FFL of the

different materials are presented in Table 2.

3.1. Fracture strengths

Fig. 1 displays the results listed by increasing scale

parameters. Regarding Weibull statistics, the fine particle

hybrid composites Filtekw Z250 and Charismaw pointed out

the significantly highest strength values while most materials

were medium ranged. The most homogenous behavior and

the lowest scatter in strength, expressed by a high Weibull

modulus m, was measured for the fine particle hybrid

composites Tetricw Ceram and Filtekw Z250. Solitairew,

however, exhibited the lowest results in both cases.

3.2. Flexural fatigue limits

In the case of cyclic fatigue measurement the strength

ranking changed. The materials with high initial strength

values point out a rather low fatigue resistance. The FFL

percentage points out a decrease in strength between 37 and

Table 2

Four-point-bending strength and fatigue strength data (SD)

Brand name Scale parameter s0 (MPa) Weibull modulus m FFL (SD) (MPa) FFL decrease (%)

Charismaw 97.94a,b 9.2 33.3 (6.2)a 65.9

Definitew 88.92b.d 9.1 47.2 (2.3)b,c 46.9

Filtekw Z250 105.16a 10.8 45.9 (7.0)b,c 56.3

Heliomolarw 91.52b 8.1 39.4 (9.1)a,c 56.9

Solitairew 55.39c 5.6 17.9 (5.1) 67.6

SolitairewII 66.80c,d 9.6 34.6 (3.8)a 48.2

Surefilw 88.59b,d 8.4 55.5 (7.2) 37.3

Tetricw Ceram 78.04b,d 12.3 45.3 (11.8)b 41.9

Data with same superscript letter are not significantly different (Mann–Whitney U-Test; a , 0.00625).

Fig. 1. Weibull plots of the investigated materials.

U. Lohbauer et al. / Dental Materials 19 (2003) 435–440 437

68%. The significantly best fatigue resistance was found for

the fine particle hybrid composite Surefilw. Here, an initial

FS of 88.59 MPa was measured and decreased within 10,000

cycles to a value of 55.5 MPa. The significantly worst fatigue

resistance was documented for the porous silica hybrid

composite Solitairew. The FS of 55.4 MPa dropped to a FFL

value of only 17.9 MPa indicating a decrease of 68%. Fig. 2

shows the FFL for the investigated materials.

Fig. 2. FFL according to the staircase method.

Fig. 3. Typical fracture surface of an initial fracture. The arrow indicates the fracture origin (a), the typical mist and hackle regions (b), and lance

hackle mark (c).

U. Lohbauer et al. / Dental Materials 19 (2003) 435–440438

3.3. Fractographic examination

The fracture surfaces of initial strength measurements

as well as those of the fatigued specimens were compared

with each other in order to find specific characteristics

regarding their fracture mechanism. Fig. 3 exhibits a

fracture surface with typical macroscopic patterns of a

fast and inert fracture. Fig. 3 is taken from the material

Tetricw Ceram and shows a representative fracture

surface compared with all resin composite materials.

The fatigue surfaces point out a quite different macro-

scopic pattern, Fig. 4. The micrograph here is taken from

the material Surefilw and shows a smooth fracture after

8503 cycles.

4. Discussion

The results for the resin composite materials indicate a

distinct variation according to their initial fracture strength.

Considering four-point-test loading, the initially investi-

gated properties behave in a similar way to those of other

studies [15,16]. To range and assess the materials’ behavior

their reliance on strength, expressed by Weibull modulus m,

has to be considered. High strength materials with a low

modulus m may be worse than lower strength materials with

less scatter in strength.

Filtekw Z250 points out the best values for both

m-value and scale parameter s0, although following

the manufacturer’s recommendation of a half light curing

period (20 s.). Solitairew on the other hand exhibited the

worst results which may be due to a delayed initiator

system or due to the porous silica fillers itself and its

optical properties [16]. Based on worse findings for

Solitairew, the material was taken from the market and

replaced by Solitairew II.

The strength ranking within the initial measurement has

changed by determination of FFL. Materials with higher

filler contents exhibited a tendency towards improved

fatigue resistance. For the highly filled Surefilw, as a so

called ‘packable’ material, the best results under cyclic

fatigue conditions were measured. However, all materials

suffer from a decrease in strength which is derived from a

mechanical fatigue within 10,000 cycles and therefore

described by FFL percentage. Surefilw shows a decrease of

37% from an initial fracture strength of 88.6 to 55.5 MPa.

This example points out that the material with the highest

initial strength value may not obviously be recommended

when focused on fatigue resistance. A correlation of FFL

with different filler types (Table 1) could not be computed.

Htang et al. [17] described a correlation of filler content on

fatigue resistance. A maximum fatigue resistance, however,

was determined with a 75 wt% filler fraction. The authors

summarized by fractographic analysis that crack propa-

gation in dental resin composites is mainly determined

through the matrix and its adhesion to the filler particles.

Drummond [8] stated, whether crack propagation in resin

composite materials is mainly around or through the second

phase particles (inter- or intracrystalline) is dependent on

filler content and interparticle distance, correspondingly.

Fig. 4. Typical fracture surface of a fatigue fracture.

U. Lohbauer et al. / Dental Materials 19 (2003) 435–440 439

This topic has to be considered to fully understand the

mechanisms of fatigue. Especially when a destructive

corrosion, caused by water exposure, weakens the matrix–

filler interface [6,7]. Ferracane et al. [18] discussed a

significant influence of silanization agents on mechanical

properties during long term water storage. To show whether

there is a correlation of different filler types and fatigue

resistance or not, the specific surface areas, particle shape

and interparticle spacing of the fillers might be compared

[8]. Hence, porous silica fillers exhibit a much larger surface

area to comparable filler sizes. Obviously extended interface

areas might be more sensitive to corrosive attack.

Further differences between FS and FFL were deter-

mined by SEM examination of the specimens’ fracture

surfaces. Figs. 3 and 4 show typical fracture surfaces for

initial and fatigue loading. The specimens were tilted in

the SEM to display the fracture origin at the top of the

micrograph. The experimental loading was applied at the

bottom (compressive zone). Fig. 3 exhibits the features

typical of a brittle fracture. The source of failure may be

located close to the tensile surface (Fig. 3(a)) surrounded by

a smooth mirror, a mist, and a hackle region in Fig. 3(b)

symmetrically around the fracture origin [19,20]. A further

crack deviating lance hackle mark is observed (Fig. 3(c)).

These marks are often determined on the compressive side

of a flexure specimen due to the obvious presence of mixed

mode conditions [21].

The failure characteristics under cyclic fatigue are

different. No sign indicates brittle fracture. The fracto-

graphic analysis shows a smooth fracture surface (Fig. 4).

This might be a hint for a diverging fracture mechanism.

Subcritical crack growth, especially under water exposure,

or maybe a moderate association to visco-elastic creep are

reported as failure criteria under cyclic fatigue loading

[22–24].

Whether the fracture mechanisms under fatigue are based

on subcritical crack growth, on visco-elastic creep or on a

combination of both, has to be cleared up in further studies.

Research also has to be done assessing the influence of

different filler types and varying specific surface areas on

fatigue behavior.

References

[1] Kohler B, Rasmusson C-G, Odman P. A five-year clinical evaluation

of class II composite resin restorations. J Dent 2000;28:111–6.

[2] Klinge S, Schmidt G, Frankenberger R, Pelka M. Clinical behavior

of solitaire: one-year results. J Dent Res 2000;97:185. Abstr. No.

332.

[3] Burke FJT, Wilson NHF, Cheung SW, Mjor IA. Influence of patient

factors on age of restorations at failure and reasons for their placement

and replacement. J Dent 2001;29:317–24.

[4] Braem M, Lambrechts P, Vanherle G. Clinical relevance of laboratory

fatigue studies. J Dent 1994;22:97–102.

[5] Freiman SW, Murville DR, Mast PW. Crack propagation studies in

brittle materials. J Mater Sci 1973;8:1527–33.

[6] Soderholm KJM, Roberts MJ. Influence of water exposure on the

tensile strength of composites. J Dent Res 1990;69:1812–6.

[7] Braem MJA, Lambrechts P, Gladys S, Vanherle G. In vitro fatigue

behavior of restorative composites and glass ionomers. Dent Mater

1995;11:137–41.

[8] Drummond JL. Cyclic fatigue of composite restorative materials.

J Oral Rehabil 1989;16:509–20.

[9] Dixon WJ, Mood AM. A method for obtaining and analyzing

sensitivity data. J Am Stat Assoc 1948;109:109–26.

[10] Draughn RA. Compressive fatigue limits of composite restorative

materials. J Dent Res 1979;58:1093–6.

[11] Soltesz U, Benkeser G. Fatigue behavior of filling materials.

Amsterdam: Oral Implantol Biomater 1989;281–6.

[12] Dieter GE. Mechanical metallurgy. London: McGraw-Hill; 1988.

[13] Weibull W. A statistical distribution function of wide applicability.

J Appl Mech 1951;18:293–8.

[14] Drummond JL, Mieschke KJ. Weibull models for the statistical

analysis of dental composite data: aged in physiologic media and

cyclic-fatigued. Dent Mater 1991;7:25–9.

[15] Manhart J, Kunzelmann KH, Chen HY, Hickel R. Mechanical

properties of new composite restorative materials. J Biomed Mater

Res (Appl Biomater) 2000;53:353–61.

[16] Choi KK, Ferracane JL, Hilton TJ, Charlton D. Properties of packable

dental composites. J Esth Dent 2000;12:216–26.

[17] Htang A, Ohsawa M, Matsumoto H. Fatigue resistance of composite

restorations: effect of filler content. Dent Mater 1995;11:7–13.

[18] Ferracane JL, Berge HX, Condon JR. In vitro aging of

dental composites in water—effect of degree on conversion, filler

volume, and filler/matrix coupling. J Biomed Mater Res 1998;42:

465–72.

[19] Frechette VD. Failure analysis of brittle materials. Advances in

ceramics, vol. 28. Westerville: American Ceramic Society; 1990.

[20] Mecholsky JJ. Fracture mechanics principles. Dent Mater 1995;11:

111–2.

[21] Mecholsky JJ. Fractography: determining the sites of fracture

initiation. Dent Mater 1995;11:113–6.

[22] Braem MJA, Davidson CL, Lambrechts P, Vanherle G. In vitro

flexural fatigue limits of dental composites. J Biomed Mater Res

1994;28:1397–402.

[23] Cock DJ, Watts DC. Time-dependent deformation of

composite restorative materials in compression. J Dent Res

1985;64:147–50.

[24] Watts DC. Elastic moduli and visco-elastic relaxation. J Dent 1994;

22:154–8.

U. Lohbauer et al. / Dental Materials 19 (2003) 435–440440