d e n t a l m a t e r i a l s 2 5 ( 2 0 0 9 ) 736–743

avai lab le at www.sc iencedi rec t .com

journa l homepage: www. int l .e lsev ierhea l th .com/ journa ls /dema

Effect of ion exchange on strength and slow crack growth ofa dental porcelain

Vinicius Rosaa, Humberto N. Yoshimurab, Marcelo M. Pintoa, Catia Fredericci b,Paulo F. Cesara,∗

a Department of Biomaterials and Oral Biochemistry, School of Dentistry, University of São Paulo, São Paulo, Brazilb Laboratory of Metallurgy and Ceramic Materials, Institute for Technological Research of the State of São Paulo, São Paulo, Brazil

a r t i c l e i n f o

Article history:

Received 2 July 2008

Received in revised form

19 September 2008

Accepted 17 December 2008

Keywords:

Dental porcelain

Ion exchange

Strength

Slow crack growth

Saliva

a b s t r a c t

Objectives. To determine the effect of ion exchange on slow crack growth (SCG) parame-

ters (n, stress corrosion susceptibility coefficient, and �f0, scaling parameter) and Weibull

parameters (m, Weibull modulus, and �0, characteristic strength) of a dental porcelain.

Methods. 160 porcelain discs were fabricated according to manufacturer’s instructions, pol-

ished through 1 �m and divided into two groups: GC (control) and GI (submitted to an ion

exchange procedure using a KNO3 paste at 470 ◦C for 15 min). SCG parameters were deter-

mined by biaxial flexural strength test in artificial saliva at 37 ◦C using five constant stress

rates (n = 10). 20 specimens of each group were tested at 1 MPa/s to determine Weibull param-

eters. The SPT diagram was constructed using the least-squares fit of the strength data

versus probability of failure.

Results. Mean values of m and �0 (95% confidence interval), n and �f0 (standard deviation)

were, respectively: 13.8 (10.1 − 18.8) and 60.4 (58.5 − 62.2), 24.1 (2.5) and 58.1 (0.01) for GC and

7.4 (5.3 − 10.0) and 136.8 (129.1 − 144.7), 36.7 (7.3) and 127.9 (0.01) for GI. Fracture stresses

(MPa) calculated using the SPT diagram for lifetimes of 1 day, 1 year and 10 years (at a 5%

failure probability) were, respectively, 31.8, 24.9 and 22.7 for GC and 71.2, 60.6 and 56.9 for

GI.

Significance. For the porcelain tested, the ion exchange process improved strength and resis-

tance to SCG, however, the material’s reliability decreased. The predicted fracture stress at

5% failure probability for a lifetime of 10 years was also higher for the ion treated group.

emy

© 2009 Acad

1. Introduction

Ceramic restorations are widely used in restorative dentistrydue to their capability to mimic the dental structure. Moreover,these materials have excellent biocompatibility, low thermalconductivity and good color stability. In spite of these qual-

∗ Corresponding author. Departamento de Biomateriais e Bioquímica OProf. Lineu Prestes, 2227 – Cidade Universitária “Armando Salles de Olifax: +55 11 55613042.

E-mail address: paulofc@usp.br (P.F. Cesar).0109-5641/$ – see front matter © 2009 Academy of Dental Materials. Pudoi:10.1016/j.dental.2008.12.009

of Dental Materials. Published by Elsevier Ltd. All rights reserved.

ities, unwanted fracture rates have been reported in clinicaltrials for porcelain inlays, onlays and crowns due to the brittle

ral – Faculdade de Odontologia – Universidade de São Paulo, Av.veira”, São Paulo – SP – Brasil. Tel.: +55 11 55613042;

nature of ceramic materials [1,2].The mechanical strength of ceramic materials depends

mainly on the presence of flaws, especially the ones locatedon the surface, since they act as stress concentrators when

blished by Elsevier Ltd. All rights reserved.

2 5

ttgiftcrmi

pitkcmepptmdsPrttpt[ret1trlashutt[cimshs

icoepctt

d e n t a l m a t e r i a l s

he material is loaded. When the stress intensity at the crackip is greater than the material’s resistance to crack propa-ation, catastrophic fracture occurs. Fracture toughness (KIc)s the material property that relates the critical flaw size withracture strength of a material. This property is directly relatedo the amount of energy the material can absorb before fastrack propagation occurs. In general, dental porcelains showelatively low values of KIc [3], especially when compared to

etals. Therefore, improving porcelain’s fracture toughnesss the key to increase its lifetime service in the oral cavity.

Various methods may be used to increase the mechanicalroperties and avoid fast crack propagation in dental ceram-

cs, such as the addition of second phase particles, thermalempering and chemical strengthening [4]. The latter, alsonown as the ion exchange process, consists of changing thehemical composition of a superficial layer of the material,aking it different from the bulk composition, by means of

xchanging smaller sodium ions (1.90 Å in diameter) for largerotassium ions (2.66 Å in diameter) from a molten salt, at tem-eratures higher than 328 ◦C (melting point of KNO3) but lowerhan the material’s strain point [5–7]. The resultant increased

olar volume in the modified surface layer leads to a two-imensional state of compression, since the expansion of theurface structure is restrained by the underlying bulk material.revious studies have shown that these compressive stressesesult in an increase in both flexural strength and fractureoughness [6,8]. The advantages of this technique over thermalempering are the relatively high magnitude of surface com-ression, the non-existence of optical distortion and the facthat it can be done in thin restorations with complex geometry5]. Moreover, the compressive layer is not lost in moist envi-onments, so improvements in the mechanical properties arexpected to be stable in the long term [9]. However, it is impor-ant to note that the exchange depth is relatively small (up to00 �m [10]) and the effectiveness of the method depends onhe amount and mobility of sodium ions present in the mate-ial. In the conventional chemical strengthening method, aong soaking time is applied to enhance both the strengthnd exchange depth. For commercial soda-lime silica glass,oaking times up to 24 h (at ∼450 ◦C in molten KNO3 bath)ave been used [11]. Alternatively, the so-called paste methodses shorter treatment times (up to 30 min) to strengthen den-al porcelains. In this method, the KNO3 paste is applied onhe surface of porcelain restorations before heat treatment8,10,12]. One disadvantage of this method is the fact thathemical strengthening may increase the fracture energy thats released during fracture of the restoration, leading to a

ultiple crack branching process that generates progressivelymaller fragments further from the fracture origin. No reportsave been found to explain if these multiple fragments repre-ent a clinical problem.

Another important property that may be affected by theon exchange process is the material’s resistance to the slowrack growth (SCG) phenomenon. SCG is the result of corrosionf ceramics metallic oxides by water molecules when pre-xisting superficial cracks are under tensile stress [13]. This

henomenon is responsible for the strength degradation oferamics over time. The dynamic fatigue method is an indirectest used to characterize slow crack growth parameters. Thisest consists of determining the fracture strength as a function

( 2 0 0 9 ) 736–743 737

of the stress rate applied in flexural strength tests. Based onthe results, it is possible to determine the fatigue parametersn (stress corrosion susceptibility coefficient) and �f0 (scal-ing parameter). The stress corrosion coefficient indicates thesusceptibility of the material to the slow crack growth phe-nomenon. The higher the n value, the lower the susceptibilityto SCG. The scaling parameter �f0 is an indication of the earlystrength of the material. The results from the dynamic fatiguetest may also be used to construct lifetime curves to predict thematerial’s strength after long periods under constant stress(e.g. 10 years) [14]. A previous study reported an increase of124% in the n value of a heat-pressed leucite-reinforced dentalglass-ceramic tested in air after a conventional ion exchangetreatment (11 h at 450 ◦C in KNO3) [6]. Nonetheless, the poten-tial of this process to increase the resistance to slow crackgrowth in moist environments still needs to be determined,since it has been shown that for dental porcelains, the n valueis usually lower in artificial saliva at 37 ◦C than in air at roomtemperature [15]. Moreover, there is a lack of information onthe effect of a moist environment in the long-term strengthbehavior of chemically strengthened dental porcelains pre-pared by the paste method.

The objective of this study was to determine the influ-ence of the ion exchange process on the strength and SCGparameters of leucite-reinforced feldspathic porcelain in arti-ficial saliva at 37 ◦C. In addition, the strength degradation overtime will be assessed by the analysis of a strength-probability-time (SPT) diagram. The hypothesis to be tested is that theion exchange process will improve the strength and the SCGparameters of the porcelain tested.

2. Material and methods

All specimens used in this study were fabricated by one singleoperator, mixing 1 g of a leucite-reinforced porcelain pow-der (Ultropaline Super Transparent, Jen Dental, Kiev, Ukraine,batch: 1094) with 0.4 mL of deionized water in order toform a slurry that was poured into a metal mold to obtaindisk-shaped specimens 15 mm in diameter and 3 mm in thick-ness. Specimens were vacuum-fired in a porcelain furnace(Keramat I, Knebel, Porto Alegre, Brazil) according to manufac-turer’s instructions. After sintering, the specimens measuring12.5 mm in diameter and 2.4 mm in thickness were machinedin a surface-grinding device (MSG-600, Mitutoyo, São Paulo,Brazil) following the guidelines in ASTM C 1161 [16] to obtainparallel surfaces and reduce thickness to 1.3 mm. Then, one oftheir surfaces was mirror-polished using a polishing machine(Ecomet 2, Buehler, Lake Bluff, USA) with diamond suspen-sions up to 1 �m to obtain the final thickness of 1.0 (± 0.1)mm.

The ion exchange strengthening treatment of choice wasthe paste method. The paste was prepared by mixing 10 g ofreagent grade KNO3 salt (Merck, Darmstadt, Germany) with4 mL of deionized water; 0.4 g of this paste was deposited onthe polished surface of each specimen and then heat treated

in an electric oven (FP-32, Yamato, Tokyo, Japan) at a heatingrate of 5 ◦C/min, an intermediary step at 150 ◦C for 20 min forpaste drying, and a holding time of 15 min at the maximumtemperature to promote the ion exchange process. A prelimi-

738 d e n t a l m a t e r i a l s 2

Table 1 – Mean and standard-deviations of biaxialflexural strength (�f), as a function of ion exchangetemperature.

Group �f (MPa)

Control 57.6 (8.9)b

I (450 ◦C) 124.2 (23.4)a

II (470 ◦C) 132.2 (10.9)a

III (490 ◦C) 136.8 (12.0)a

Values followed by the same superscript are statistically similar(p > 0.05).

nary study was carried out to determine the ideal temperaturefor ion exchange in the porcelain studied. Four groups (n = 10)were prepared according to the temperature used: I (control,no ion exchange); II (450 ◦C); III (470 ◦C); and IV (490 ◦C). Theproperty determined was flexural strength, according to themethodology described below. These results were analyzed byone-way analysis of variance (ANOVA) and Tukey’s test with aglobal significance level of 0.05.

To determine the influence of ion exchange on strength andSCG parameters, 160 disk-shaped specimens were producedand 80 of them were subjected to the ion exchange process at470 ◦C. This temperature was chosen because no significantdifferences were observed in the flexural strength as a func-tion of ion exchange temperature (Table 1). Moreover, the workby Zijlstra and Burggraaf [17] suggested an ion exchange treat-ment temperature 100 ◦C below the material’s glass transitiontemperature, which was 575 ◦C for the porcelain used in thepresent study.

The biaxial flexure strength (�f) was determined using thepiston-on-three-balls method (ASTM F394-78 [18]) in a univer-sal testing machine (Syntech 5G, MTS, São Paulo, Brazil) atconstant stress rates with the specimen immersed in artifi-cial saliva [19] heated and maintained constant at 37 ◦C. Thebiaxial flexure strength was calculated using the following Eq.[20]:

�f = 3F(1 + �)4�t2

[1 + 2 ln

Ra

b+ 1 − �

1 + �

(1 − b2

2R2a

)R2

a

R2

](1)

where F is the fracture load, t is the disk thickness, � is Pois-son’s ratio (0.22), b is the radius of the loading area (pistonradius, 0.85 mm), R is the disk radius, and Ra is the radius of thesupport circle defined by the three balls (4.0 mm). The valueof � was determined by the ultrasonic pulse-echo method[21], using a 200 MHz ultrasonic pulser-receiver (Panametrics,USA, 5900 PR), 20 MHz longitudinal and shear transducerswith a delay material, and a coupling paste applied betweenthe sample and transducer [22]. No significant differenceswere observed between the � values of the two experimentalgroups. In the preliminary study a fixed stress rate (10 MPa/s)was used to evaluate the effect of ion exchange temperatureon the flexural strength. The SCG parameters were evaluatedby the dynamic fatigue test, and specimens were tested in

biaxial flexure at each of five stress rates: 10−2, 10−1

, 100, 101

and 102 MPa/s (n = 30 at 1 MPa/s and n = 10 at the remainingrates), immersed in artificial saliva (at 37 ◦C) with the followingcomposition: 100 mL of KH2PO4 (2.5 mM); 100 mL of Na2HPO4

5 ( 2 0 0 9 ) 736–743

(2.4 mM); 100 mL of KHCO3 (1.5 mM); 100 mL of NaCl (1.0 mM);100 mL of MgCl2 (0.15 mM); 100 mL of CaCl2 (1.5 mM); and 6 mLof citric acid (0.002 mM). The parameters n (stress corrosionsusceptibility coefficient) and �f0 (scaling parameter) were cal-culated according to the following Eq. [23]:

log �f =(

1n + 1

log �̇ + log �f0

)(2)

where �̇ is the stress rate.Strength results of the specimens tested at 1 MPa/s (n = 30)

were fitted to the two-parameter Weibull distribution to cal-culate the strength parameters m (Weibull modulus) and �0

(characteristic strength), based on the following Eq. [24]:

Pf = 1 − exp[−(

�f

�0

)m](3)

where Pf is the fracture probability. The parameter �0 corre-sponds to the strength at the failure probability of 63.2%. Basedon Eq. (3), the parameter �5% was also determined, whichcorresponds to the strength at the more clinically relevantfailure probability of 5%. The Weibull parameters were cal-culated based on the maximum likelihood method, accordingto the ASTM C 1239 [25] and plotted in addition to their 95%confidence level and the 95% confidence interval of Weibullmodulus matched to characteristic strength.

The introduction of the Weibull statistics and the timedependency of strength makes it possible to build strength-probability-time (SPT) diagrams, where the time to failureunder constant stress for different failure probability levelscan be estimated using the results obtained in the dynamicfatigue test, according to the theory proposed by Davidge etal. [26]. In the present study, the SPT diagram matches thefailure probabilities and the applied constant stress to threedifferent times to failure: 1 day, 1 year and 10 years.

Fracture surfaces of selected specimens broken in the biax-ial flexure test were examined using optical microscopy (OM)and scanning electron microscopy (SEM) to investigate thefracture origin. The critical flaws were identified and theirsizes were determined based on fractographic principles [27].Critical crack size (c’) was calculated as follows:

c′ = (a′b′)1/2 (4)

where a′ is the crack depth and b’ is its half crack width. Forthis analysis, only three stress rates were considered: 0.01; 1and 100 MPa/s (n = 4), since the objective was to observe theoccurrence of slow crack growth by comparing critical cracksizes at different stress rates. The specimens selected for thisanalysis were those with strength as close as possible to thecalculated mean value of their groups.

3. Results

Results of the preliminary study, i.e. mean values of flexural

strength as a function of different ion exchange temperaturesare listed in Table 1. A significant increase in flexural strengthcan be observed after the ion exchange treatment. Moreover,it is possible to note that different ion exchange temperatures

d e n t a l m a t e r i a l s 2 5 ( 2 0 0 9 ) 736–743 739

Table 2 – Weibull parameters (95% confidence intervals in parentheses).

Group m �0 (MPa) �5% (MPa) n �f0 (MPa) �10 (MPa)

Control 13.8 (10.1–18.8) 60.4 (58.5–62.2) 48.7 (43.4–53.0) 24.1 (2.5) 58.1 (0.01) 31.1Ion exchange 7.4 (5.3–10.0) 136.8 (129.1–144.7) 91.6 (73.7–107.8) 36.7 (7.3) 127.9 (0.01) 90.4

m: Weibull modulus, �0: characteristic strength, �5%: strength at 5% failure probability, and fatigue parameters (standard deviations), n: stresscorrosion coefficient; �f0: scaling parameter; �10: predicted strength after 10 years, for control and ion treated group.

Fig. 1 – Weibull plot showing the results of the flexuralstrength test (1 MPa/s, saliva, 37 ◦C) for both experimentalgroups. Dashed lines represent the 95% confidence intervalof Weibull modulus and straight lines represent the 95%confidence interval of the Weibull modulus combined witht

r

stbetctfdhgb8dev

r(cc

Fig. 2 – Flexural strength as a function of stress rate for

sponding to a time to failure of 10 years (�10) were calculatedusing the equation obtained in the regression analysis shownin Fig. 3. The �10 values obtained were 90.4 and 31.1 MPa for theion exchange group and control group, respectively (Table 2).

he characteristic strength.

esulted in similar flexural strength for the porcelain tested.Table 2 shows the Weibull parameters (�0, �5% and m),

ubcritical crack growth parameters (n and �f0), and the frac-ure stress corresponding to a lifetime of 10 years (�10) foroth groups (control and ion exchanged). Although the ionxchange increased the characteristic strength (�0) in 126%,his process caused a 46% decrease in the m value. The 95%onfidence intervals for both �0 and m values indicate thathese differences are statistically significant as the intervalsailed to overlap. The Weibull plot for both experimental con-itions is presented in Fig. 1, where it is possible to note theigher variability of the strength data in the ion exchangeroup. In addition, it can be noted that at the failure proba-ility of 5%, the strength of the ion exchange group (�5%) was8% greater than that of the control. However, if the 95% confi-ence interval of �5% is considered, the lowest value of the ionxchange group (73.7 MPa) is only 40% higher than the highestalue of the control group (53.0 MPa) (Table 2).

The results of biaxial flexure strength at different stress

ates are shown in Fig. 2. Both slow crack growth parametersn and �f0) increased significantly after the ion exchange pro-ess (Table 2). Fig. 3 represents the lifetime curves for bothonditions. The slopes of these curves are directly related to

control and ion exchange groups measured in artificialsaliva at 37 ◦C.

the material’s n value (the higher the n value, the lower theslope). Thus, the lower n value of the control group comparedto ion exchanged material resulted in the higher slope of thefirst, denoting that the ion treatment increased in ∼50% then value of the porcelain tested. The fracture stresses corre-

Fig. 3 – Plots of the flexural strength as a function of time tofailure for both experimental groups. The time axis islabeled for l day (1d), l year (1y) and 10 years (10y).

740 d e n t a l m a t e r i a l s 2 5 ( 2 0 0 9 ) 736–743

Fig. 4 – SPT (strength-probability-time) diagram for bothexperimental groups.

Table 3 – Fracture stresses (MPa) corresponding to 5%failure probability at different times for bothexperimental groups. Values obtained from the SPTdiagram (Fig. 4).

Group Initial 1 day 1 year 10 years

Fig. 5 – Representative SEM image of a fracture surface of

increase in a material’s reliability [31,32], although no signif-icant changes [6] or a decrease in the Weibull modulus [33]after the treatment have also been reported. In the case ofion exchange treatments that use shorter times, the limited

Control 49.6 31.8 24.9 22.7Ion exchange 94.6 71.2 60.6 56.9

The SPT diagram obtained from the dynamic fatigue dataat different times to failure is shown in Fig. 4. It is observedthat strength degradation over time is higher for the controlgroup as compared to the ion exchange group. Considering the5% failure probability, the corresponding fracture stresses forthe times to failure of 1 day, 1 year and 10 years are shown inTable 3. After 10 years, the predicted decrease in fracture stressat such failure probability is approximately 40% and 54% forthe treated and control group, respectively.

Fractography revealed that the fracture origins in all eval-uated specimens were surface flaws of semi-elliptical shape,located in the polished surface (Fig. 5). Mean critical crack sizesas a function of the stress rate are shown in Fig. 6. It is possibleto note a decrease in the critical crack size with the increasein stress rate in both experimental groups. Moreover, all cracksizes measured were higher in the control group, regardless ofthe stress rate.

4. Discussion

The hypothesis of this work was accepted since the resultsshowed that the ion exchange process improved the flexuralstrength and the SCG parameters of the porcelain tested. The

increase in flexural strength observed after the ion exchangeprocess in the preliminary results (Table 1) was expectedbecause of the development of a compressive layer in thematerial’s surface due to the replacement of small Na+ ions

one ion exchange specimen fractured in three pieces at1 MPa/s. The arrows indicate the crack front.

for larger K+ ions in the porcelain’s glassy matrix. Becauseof this surface modification, cracks in the porcelain will onlypropagate when the stresses generated during loading over-come the residual stresses induced by the compressive layerat the surface. These results also indicated that in the tem-perature range investigated, the strengthening effect causedby the increase in the ion diffusion rate due to the increasein temperature was compensated by stress relaxation causedby the accommodation of these ions in the glass network, bymeans of diffusion and viscous processes [28]. Previous studiesalso showed that the flexural strength of leucite-based porce-lains was not affected by temperature variations in a rangesimilar to that tested in the present study [29,30].

The ion exchange process used in the present studyaffected the Weibull parameters in different ways, sincethe characteristic strength (�0) increased and the Weibullmodulus (m) decreased significantly after the ion exchange(Table 2 and Fig. 1). In other words, although the treatmentresulted in a 126% increase in the material’s strength, itsreliability decreased significantly due to the high variabilityin the strength results (from 89 to 161 MPa). The conven-tional extended ion exchange treatments usually result in an

Fig. 6 – Critical crack size as a function of stress rate forboth experimental groups.

d e n t a l m a t e r i a l s 2 5

Fig. 7 – Residual surface stress profiles expected fromdifferent ion exchange (IE) methods: conventional extendedIE; two-step IE; and paste method (short IE treatment).Vertical dotted lines represent the depths of a small and al

esbseirpimw7imsaowtecfmwsetto

otp

arge surface crack.

xchange depth strengthens preferably the originally strongerpecimens (i.e. those with the smallest Griffith flaws), thusroadening the strength distribution [34]. In addition, con-idering that the compressive layer resulting from the ionxchange by the paste method is less than 100 �m thick [10],t is possible that while smaller defects were totally sur-ounded by the compression stresses, larger flaws were onlyartially surrounded, thereby contributing for the variabil-

ty in the results. In order to increase the reliability of glassaterials, a two-step ion-exchange method was proposed, inhich a short second ion-exchange process in a 30% Na2O-

0% K2O bath was carried out for partial removal of the K+

ons introduced in the first extended treatment [35]. Thisethod produces a stress profile with the maximum compres-

ive stress shifted to a given distance from the surface, whichcts as a barrier to the crack propagation [31]. The second stepf this method, however, tends to lower the strength achievedith the first ion-exchange step [6]. Fig. 7 shows schematically

he residual surface stress profiles expected from different ionxchange methods [6,10,35,36], with the indications of tworack sizes (small and large). It is important to note that, apartrom the noticeable increase in variability, the overall perfor-

ance of the ion exchange group in terms of flexural strengthas better than that of the control group, since the lowest

trength value measured in the first was higher than the high-st value obtained for the latter. In addition, Table 2 showshat the ion-exchange group had a significantly higher frac-ure strength at the more clinically relevant failure probabilityf 5% (�5%), considering the 95% confidence interval.

The higher stress corrosion susceptibility coefficient (n)btained after the ion exchange (Table 2 and Fig. 3) indicateshat this process not only increased the strength of the dentalorcelain studied but also reduced its degradation over time

( 2 0 0 9 ) 736–743 741

in the artificial saliva environment at 37 ◦C. The ion exchangematerial showed an increased resistance against slow crackgrowth due to the presence of compressive stresses in thesurface layer. In fact, it is possible to note in the fractographicanalysis (Fig. 6) that critical crack sizes were larger for the con-trol group especially at the stress rate of 0.01 MPa/s, indicatingthat in this group the absence of residual compressive stressesallowed cracks to grow at a higher rate. The ability of the ionexchange treatment in reducing slow crack propagation canalso be noted by comparing the values of �f0 (early strength)and �10 (predicted strength after 10 years) in Table 2. Theseresults show that the degradation in strength after 10 yearsof service was higher for the control group (decreased 47%,from 58.1 to 31.1 MPa) as compared to the ion exchange treatedgroup (decreased 29%, from 127.9 to 90.4 MPa). Such differ-ence in strength degradation can also be observed in Fig. 3,in which the curve corresponding to the control group showsthe steeper slope.

Fig. 1 shows that in the low failure probability range, thelower Weibull modulus (m) of the ion-exchanged porcelainresults in flexural strength values that are closer to that of thecontrol porcelain, decreasing the difference observed between�0 values. The higher stress corrosion susceptibility coeffi-cient (n) of the ion-exchange porcelain, however, compensatedthis effect partially, since its slower strength degradationrate resulted in an increase in the difference between thestrength of both groups for any level of fracture probability(Fig. 4).

The higher strength degradation of the control group is alsoobserved in the SPT diagram in Fig. 4, since the period betweencurves is larger for this group. This graph also gives us theperspective of what would be the stress to fracture 5% of thespecimens after 10 years of service (Table 3). Such stress isapproximately 100% higher for the treated group, indicatingthat the ion exchange leads to a significant improvement inthe porcelain service lifetime. In order to extrapolate theseresults to clinical practice, it is necessary to consider thattimes to failure described in the SPT diagram are for staticloading calculated from dynamic fatigue tests [26]. Therefore,these results may not correlate well with the cyclic loadingpattern of the mastication process. However, it is possible toestimate, for a certain time of cyclic loading regimen, the cor-responding time in static loading by adding all the periodsof effective mastication. In this way, for a clinical lifetime ofabout 10 years, one can assume that the restoration would besubjected to average masticatory stresses for only 1–4% of thetime, which corresponds to a period varying from 38 to 152days. These periods of time fall within the periods of 1 dayand 1 year of static loading presented in Table 3. Consider-ing that the mean applied stress in each chewing cycle in themolar region is approximately 28 MPa [37], it is possible to notethat for both periods of 1 day and 1 year (Table 3) the fracturestresses (at 5% failure probability) of the control group are veryclose to the stress of 28 MPa described above, while the valuesobtained for the treated group indicate a more reliable situa-tion, since they are at least double the stress of one chewing

cycle in the molar region.

The clinical parameters described above can also be appliedto the SPT diagram to calculate the failure probability of theexperimental groups taking into consideration a static load-

742 d e n t a l m a t e r i a l s 2

r

Fig. 8 – Number of fracture pieces as a function of thebiaxial flexural strength.

ing at 28 MPa for a period of 38 or 152 days. In this case, theestimated failure probabilities of the control group would varyfrom 8 to 19%, while no failures would be observed in thetreated group. Such predictions should be carefully consideredsince the magnitude, duration and stress development in theoral environment are quite complex. In this regard, it should betaken into account that the cyclic nature of mastication com-prises long periods without the application of stresses, whichmay result in crack healing or crack blunting [38]. Other fac-tors should also be evaluated such as the restoration thicknessand shape, defects introduced during processing or adjustingand the presence of residual stresses.

Fig. 8 was constructed in order to assess if there might beany clinical concern regarding the number of fracture piecesof an ion-exchange restoration fractured in the mouth. It ispossible to note that all the specimens in the control groupfractured in only two pieces, while most of the ion exchangespecimens fractured in 3 or 4 pieces. The higher number ofpieces in the ion exchange group is related to the higher frac-ture energy that is released during crack propagation. From theclinical point of view, the number of fracture pieces found inthe present work should not represent a clinical problem. Thisstatement is limited only to the porcelains strengthened bythe paste method (using short ion exchange treatment), sinceglasses strengthened by the conventional extended chemi-cal strengthening method can fragment in a large numberof small pieces, depending on the residual stress levels intro-duced during the treatment [36].

5. Conclusion

Based on the results of this study it was possible to con-clude that the ion exchange treatment using the paste methodimproved strength, reduced slow crack growth (increased thestress corrosion susceptibility coefficient, n) in artificial salivaat 37 ◦C, and reduced strength degradation over time for the

leucite-reinforced feldspathic porcelain tested. The significantincrease in strength and the higher coefficient n compensatedfor the noticeable reduction in reliability, as represented by thedecrease in the Weibull modulus after the treatment.

5 ( 2 0 0 9 ) 736–743

Acknowledgments

The authors would like to thank Dr. Victor Styopkin for pro-viding the porcelain used in this study. Based on a thesissubmitted to Universidade de São Paulo, in partial fulfilmentof the requirements for the M.Sc. degree. This research wassupported by Brazilian agencies CAPES, FAPESP and CNPq forthe financial support of the present research.

e f e r e n c e s

[1] Hayashi M, Tsuchitani Y, Kawamura Y, Miura M, Takeshige F,Ebisu S. Eight-year clinical evaluation of fired ceramic inlays.Oper Dent 2000;25:473–81.

[2] Smales RJ, Etemadi S. Survival of ceramic onlays placed withand without metal reinforcement. J Prosthet Dent2004;91:548–53.

[3] Cesar PF, Yoshimura HN, Miranda Junior WG, Okada CY.Correlation between fracture toughness and leucite contentin dental porcelains. J Dent 2005;33:721–9.

[4] Odo GY, Nogueira LN, Lepienski CL. Ionic migration effectson the mechanical properties of glass surfaces. J Non-CrysSolids 1999;247:232–6.

[5] Varshneya AK. Physical properties of ion-exchanged andmelt-processed glasses differ. Glass Researcher 2001;10:21–7.

[6] Fischer H, Marx R. Suppression of subcritical crack growth ina leucite-reinforced dental glass by ion exchange. J BiomedMater Res A 2003;66:885–9.

[7] Bartholomew RF. Ion exchange. 4 EUA: ASM International;1991.

[8] Seghi RR, Crispin BC, Mito W. The effect of ion exchange onthe flexural strenght of feldsphatic porcelains. Int JProsthodont 1990;3:130–4.

[9] Fischer H, Tinschert J, Marx R. Steigerung derBeanspruchbarkeit vollkeramischer Brücken durchIonenaustaschverfahren Dtsch Zahnarztl Z 1999;54:321–4.

[10] Piddock V, Qualtrough AJ, Brough I. An investigation of anion strengthening paste for dental porcelains. Int JProsthodont 1991;4:132–7.

[11] Sinton CW, Lacourse WC, O’Connell MJ. Variations in K+-Na+

ion exchange depth in commercial and experimental floatglass compositions. Mater Res Bull 1999;34:2351–9.

[12] Cesar PF, Gonzaga CC, Miranda Jr WG, Yoshimura HN. Effectof ion exchange on hardness and fracture toughness ofdental porcelains. J Biomed Mater Res B Appl Biomater2007;83:538–45.

[13] Ritter JE. Crack propagation in ceramics. 4 Metals Park: OH:ASM International; 1991.

[14] Wachtman JB. Mechanical properties of ceramics. New York:John Wiley & Sons; 1996.

[15] Cesar PF, Soki FN, Yoshimura HN, Gonzaga CC, Styopkin V.Influence of leucite content on slow crack growth of dentalporcelains. Dent Mater 2008;24:1114–22.

[16] ASTM. C 1161 Flexural srength of advanced ceramics atambient temperature. American Society for TestingMaterials, 2002.

[17] Zijlstra AL, Burggraaf AJ. Fracture phenomena and strenghtproperties of chemically and physically strengthened glass.II. Strength and fracture behavior of chemicallystrengthened glass in connection with the stress profile. J

Non-Crys Solids 1969;1:163–85.

[18] ASTM Designition. Standard test method for biaxial flexurestrength (modulus of rupture) of ceramics substrates.Philadelphia, PA: American Society for Testing Materials;F394–78 [reapproved 1991].

2 5

d e n t a l m a t e r i a l s

[19] Cate JMT, Duijisters PPE. Alternating demineralization andremineralization of artificial enamel lesions. Caries Res1982;16:201–10.

[20] Shetty DK, Rosenfield AR, Mcguire P, Bansal GK, DuckworthWH. Biaxial flexure tests for ceramics. Am Ceram Soc Bull1980;59:1193–7.

[21] JIS. Testing methods for elastic modulus of highperformance ceramics. JIS R 1602 1986;216.

[22] Yoshimura HN, Molisani AL, Narita NE, Cesar PF, H. G.Porosity dependence of elastic constants in aluminumnitride ceramics. Mater Res 2007;10:127–33.

[23] ASTM. C 1368-00 Standard test method for determination ofslow crack growth parameters of advanced ceramics byconstant stress-rate flexural testing at ambient temperature.Annual Book of ASTM 2001;15:1–9.

[24] Weibull W. A statistical distribution of wide applicability. JAppl Mechan 1951;18:293–7.

[25] ASTM. C1239 Reporting uniaxial strength data andestimating Weibull distribution parameters for advancedceramics. American Society for Testing Materials, 2000.

[26] Davidge RW, McLaren JR, Tappin G.Strength-probability-time (SPT) relationship in ceramics. JMater Sci 1973;8:1699–705.

[27] Morena R, Lockwood PE, Fairhurst CW. Fracture toughness of

commercial dental porcelains. Dent Mater 1986;2:58–62.

[28] Orgaz F, Navarro JMF. Prediction of results of strengtheningglass by using the misfitting sphere theory. New York:Plenum Press; 1985.

( 2 0 0 9 ) 736–743 743

[29] White SN, Seghi RR. The effect of ion strengtheningtime/temperature kinetics on the flexural strength offeldspathic porcelains. Dent Mater 1992;8:320–3.

[30] Denry IL, Rosenstiel SF, Holloway JA, Niemiec MS. Enhancedchemical strengthening of feldspathic dental porcelain. JDent Res 1993;72:1429–33.

[31] Sglavo VM, Green DJ. Flaw insensitive ion-exchanged glass:II, production and mechanical performance. J Am Ceram Soc2001;84, 1832-1238.

[32] Fischer H, Marx R. Improvement of strength parameters of aleucite-reinforced glass ceramic by dual-ion exchange. JDent Res 2001;81:336–9.

[33] Fischer H, Maier HR, Marx R. Improved reliability of leucitereinforced glass by ion exchange. Dent Mater 2000;16:120–8.

[34] Varner JR. The practical strength of glass. New York: PlenumPress; 1985.

[35] Green DJ, Tandon R, Sglavo VM. Crack arrest and multiplecracking in glass through the use of designed residual stressprofiles. Science 1999;283:1295–7.

[36] Glass SJ, Tandon R, Heller D. Characterization of crackbranching and fragmentation patterns for ion-exchangedglass. In: Ceram Transactions, v. 1999. Hoboken, New Jersey:The American Ceramic Society & John Wiley & Sons; 2007.

[37] Lohbauer U, Kramer N, Petschelt A, Frankenberger R.

Correlation of in vitro fatigue data and in vivo clinicalperformance of a glass ceramic material. Dent Mater2008;24:39–44.

[38] Fairhurst CW, Lockwood PE, Ringle RD, Twiggs SD. Dynamicfatigue of feldsphatic porcelain. Dent Mater 1993;9:269–73.