Influence of pH on slow crack growth of dental porcelains

d e n t a l m a t e r i a l s 2 4 ( 2 0 0 8 ) 814–823

available at www.sciencedirect.com

journal homepage: www.intl.elsevierhealth.com/journals/dema

Influence of pH on slow crack growth of dental porcelains Marcelo M. Pinto a , Paulo F. Cesar a,∗ , Vin´ıcius Rosa a , Humberto N. Yoshimura b a b

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

a r t i c l e

i n f o

a b s t r a c t

Article history:

Objectives. To evaluate the effect of pH of storage medium on slow crack growth (SCG) param-

Received 14 September 2007

eters of dental porcelains.

Received in revised form

Methods. Two porcelains were selected: with (UD) and without (VM7) leucite particles, in

25 September 2007

order to assess if the microstructure would affect the response of the material to the pH

Accepted 3 October 2007

variation. Disc specimens were produced following manufacturers’ instructions. Specimens were stored in artificial saliva in pHs 3.5, 7.0 or 10.0 for 10 days and after that the fatigue parameters (n: SCG susceptibility coefficient and  0 : scaling parameter) were obtained by the

Keywords:

dynamic fatigue test using the same pH of storage. Microstructural analysis of the materials

Dental porcelain

was also performed.

Slow crack growth

Results. For VM7, the values of n obtained in the different pHs were similar and varied from

Fractography

29.9 to 31.2. The  0 value obtained in pH 7.0 for VM7 was higher than that obtained in the

Microstructure

other pHs, which were similar. For porcelain UD, n values obtained in pHs 7.0 and 10.0 were

Fatigue

similar (40.8 and 39.6, respectively), and higher than that obtained in pH 3.5 (26.5). With respect to  0 , the value obtained for porcelain UD in pH 10.0 was lower than those obtained in pHs 3.5 and 7.0, which were similar. Significance. The effect of pH on the stress corrosion susceptibility (n) depended on the porcelain studied. While the n value of VM7 was not affected by the pH, UD presented lower n value in acid pH. For both porcelains, storage in acid or basic pH resulted in strength degradation. © 2007 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

1.

Introduction

Porcelains are aesthetic materials largely used in dentistry to construct varied types of restorations and prostheses. However, due to their low fracture toughness, unwanted fracture rates have been reported in clinical trials [1,2]. Therefore, the mechanical behavior of dental porcelains still needs to be improved, and consequently the understanding of its fracture behavior is key for the development of materials with longer lifetime. The fracture of ceramics in service occurs with little or no plastic deformation when small flaws or cracks propagate in

∗

an unstable manner under applied tensile stresses (i.e. catastrophic failure) [3]. For a body containing a crack of size c, subjected to a tensile stress , the stress intensity factor at the tip of the crack is given by √ KI = Y c

(1)

where Y is a dimensionless constant which depends on the stress mode, geometry and dimensions of the material, shape and length of the crack. Fracture occurs when the critical level of stress intensity factor is reached (KIc ). When subjected to a stress intensity factor below the critical level (KI < KIc ), the defects in ceramic materials may present

´ ˜ Paulo), Av. Prof. Lineu Corresponding author at: Departamento de Materiais Dentarios, Faculdade de Odontologia (Universidade de Sao ´ ˜ Paulo, SP, Brazil. Tel.: +55 11 5561 3042; Prestes, 2227 Cidade Universitaria “Armando Salles de Oliveira”, CEP 05508-900, Sao fax: +55 11 5561 3042. E-mail address: [email protected] (P.F. Cesar). 0109-5641/$ – see front matter © 2007 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.dental.2007.10.001

d e n t a l m a t e r i a l s 2 4 ( 2 0 0 8 ) 814–823

a slow and stable growth, mainly in a moist ambient environment, such as the oral environment. This phenomenon is called slow or subcritical crack growth (SCG) and leads to strength degradation over time [4]. The presence of water at the tip of a crack under stress results in the rupture of the metallic oxide bonds of the material, with the subsequent formation of hydroxides. As a consequence of stress corrosion, a defect may reach the critical size (under a determined applied stress) and result in fast fracture [5]. The subcritical crack growth is notable for its extreme sensitivity to applied stress and it tends also to depend on concentration of environmental species, temperature and other extraneous variables [6]. Crack velocity, v, during slow crack propagation may be described by the following power law [6]: v = AKIn

(2)

where A and n are constants dependent on the material properties and environmental variables. Some aspects of this relationship are well documented, such as three distinct regions observable in the v–KI plot, designated I, II, and III, each with its own characteristic. Region I has attracted the most attention in literature because it is highly dependent on KI and environmental conditions [7]. Dental prostheses are usually subjected to stress levels corresponding to this region during most part of their lifetimes. In region I, the environmental species react with the ceramic interatomic bonds in the crack front, leading to stable crack propagation [6]. Region II is associated with intermediate crack velocities and also depends on environment, but is much less sensitive to KI [8]. This region depends on the diffusion of the corrosive species from the environment to the crack tip. In Region III, ultrasonic velocities are reached, and crack growth is strongly dependent on KI , but is environment-insensitive, since it is usually associated with the kinetic of intrinsic bond rupture [9]. The oral environment presents many elements that favor SCG in ceramic restorations, such as (a) water from saliva; (b) water from the luting cement and from the dentin tubules; (c) masticatory stresses; (d) stresses associated to differences in the coefficient of thermal expansion of the restoration components; (e) temperature variations; and (f) pH variations [10]. In this way, the slow growth of defects is expected to be a common finding in porcelain restorations, resulting in a decrease of their strength and in-service reliability [11]. With regard to pH variations, it is important to consider that porcelain restorations are immersed in saliva in the oral cavity, and subjected to periodic pH alterations [12–14]. Though the normal pH of saliva varies from 6.8 to 7.2 [15], when an individual takes carbohydrates into the mouth the dental plaque produces organic acids, resulting in acidic pH (around 4.5). As a response to this phenomenon, the saliva flux is increased in order to provide bicarbonate ions that will increase the oral pH to the normal level again (buffering capacity). The oral pH can also be altered directly by the type of food ingested. Thus, acidic products such as lemonade and soft drinks lead to a decrease of the oral pH without the involvement of microorganisms.

815

Some works investigated the effect of the pH of the storage medium on crack growth of glasses and demonstrated that crack velocity usually increases with the increase in the pH of the storage medium [16,17]. Another work also demonstrated that alkaline solutions containing NaOH dissolve the surface of glasses, leading to slow crack growth and consequently degrading the material’s strength [18]. On the other hand, it was also demonstrated that for soda-lime glass crack velocity may not vary as function of pH [19], and in some cases it may even increase in acid pH [16]. The effect of pH variations on slow crack growth of dental porcelains still needs to be determined. This information is important for clinicians and manufacturers since it may help the tailoring of the porcelain’s microstructure in order produce materials with lower strength degradation in the oral cavity over time. The objective of the present study was to evaluate the effect of pH of storage medium on slow crack growth parameters of dental porcelains. Two porcelains were selected (with and without leucite particles) in order to assess if the microstructure would affect the response of the material to the pH variation. The main hypothesis to be tested is that the pH will affect the slow crack growth parameters of both materials.

2.

Materials and methods

2.1.

Specimen preparation

Two dental porcelain powders were used: (a) Veneer Material 7—VM7 (Vita Zahnfabrik, Germany); and (b) Ultropaline Dentin—UD (Jendental, Ukraine). VM7 is a vitreous porcelain indicated as veneering material for glass-infiltrated alumina cores, and UD is a leucite-based porcelain indicated for porcelain-fused-to-metal. Both materials can also be used in all-ceramic restorations. These porcelains were chosen in order to investigate if the microstructure (with and without leucite) influences the response of the material to the pH variation. The green specimens (14.9 mm in diameter and 2.9 mm in thickness) were prepared by the vibration–condensation method and sintered in a dental porcelain furnace (Keramat I, Knebel, Porto Alegre, Brazil) following the firing schedules recommended by the manufacturers. After firing, the specimens were machined to achieve thickness of 1.3 mm, following the guidelines in ASTM C 1161 [20] (the final diameter of the specimen was 13.0 mm). Then, one of the plain surfaces was mirror polished to the final thickness of 1.00 (±0.13) mm, using a polishing machine (Ecomet 3, Buehler, Lake Bluff, IL, USA) with diamond suspensions (45, 15, 6 and 1 ␮m). A total of 150 specimens were produced for each porcelain.

2.2.

Dynamic fatigue testing

Before testing, three solutions were produced with pH of 3.5, 7.0 and 10.0. These solutions had the following composition: 100 mL of KH2 PO4 (2.5 mM); 100 mL of Na2 HPO4 (2.4 mM); 100 mL of KHCO3 (1.50 mM); 100 mL of NaCl (1.0 mM); 100 mL of MgCl2 (0.15 mM); 100 mL of CaCl2 (1.5 mM); and 6 mL of citric acid (0.002 mM). The pH of the solution was adjusted to 3.5 by

816

d e n t a l m a t e r i a l s 2 4 ( 2 0 0 8 ) 814–823

adding HCl, and to 10.0 by adding NaOH. The 150 specimens of each material were equally divided into the three solutions and stored for 10 days at 37 ◦ C. After the storage period, the dynamic fatigue test was carried out based on the ASTM C 1368-00 [21]. The specimens were fractured in biaxial flexure by means of the pistonon-3-balls technique in a universal testing machine (MTS ˜ Paulo, Brazil). The disks were supported on Syntech 5G, Sao three hardened steel balls (1.6 mm in diameter) that were evenly spaced around a support circle having a radius of 4 mm. The specimens were loaded to failure by a flexural load applied by a steel cylinder with a diameter of 1.6 mm. The fracture stress ( f ) was calculated according to the following formula:



f =

3P(1 + v) a (1 − v) 1 + 2 ln + b (1 + v) 4t2

 1−

b2 2a2



a2 R2

 (3)

where P is the fracture load, t the specimen’s thickness, a the radius of the support circle, b the radius of the steel cylinder, v the Poisson’s ratio (determined by the pulse-echo technique) and R is the specimen’s radius. The biaxial flexural tests were performed with specimens immersed at 37 ◦ C in a glass box with the same solution in which they were stored for 10 days. To determine the dynamic fatigue parameters, 10 specimens of each material were tested at one of the following stress rates: 10−2 , 10−1 ,100 , 101 and 102 MPa/s. Fatigue parameters (n and  0 ) and their standard deviations were calculated according to the equations presented in the ASTM C 1368-00 [21], which are based on the following equation: log  =

1 log ˙ + log 0 n+1

(4)

where  is the flexural strength, ˙ the stress rate, n the stress corrosion susceptibility coefficient and  0 is a scaling parameter. Fracture surfaces of selected specimens were examined using optical microscopy (OM, Leica DMRXE, Bensheim, Germany) and scanning electron microscopy (SEM, Jeol JSM 6300, Peabody, USA) to investigate the fracture origin. The critical flaws were identified and their size was determined based on fractographic principles [22]. Critical crack size (c) was calculated as follows: c = (ab)

1/2

 E 0.5  P

KIc = 0.016

H

cr1.5

(6)

where P is the indentation load, cr the average size of the median/radial cracks, E the elastic modulus (determined by the pulse-echo technique) and H is the material’s hardness (H = 0.5 P/˛2 , where ˛ is the average half-diagonal of the indentation).

2.4.

Microstructural analysis

Microstructural analysis was made after etching the polished surfaces of both porcelains with 2% hydrofluoric acid (HF) solution for 15 s in order to reveal the microstructure. For the leucite-based porcelain, 10 images of different surface areas were obtained at the scanning electron microscope (SEM). After improving contrast and brightness with specific software (Adobe Photoshop 7.0® , Seattle, USA), images were taken into an image analyzer software (Leica QWin, Selb, Germany) to determine the leucite content in vol%, and mean particle size. X-ray diffraction (XRD) was performed taking powders of each material into a diffractometer (Rigaku Rint 2000, Tokyo, Japan), using the Cu K␣ radiation and scan rate of 1◦ of 2 per minute. Semi-quantitative chemical analysis of both materials was also performed by means of X-ray fluorescence spectroscopy (XRF 1500, Shimadzu, Japan).

2.5.

Statistical analysis

Biaxial strength, fracture toughness and hardness data were analyzed by one-way ANOVA and multiple comparisons were performed using Tukey’s post hoc test at a pre-set significance level of 5%. The fatigue parameters were analyzed according to the guidelines presented in the ASTM C 1368-00 [21]. Lifetime curves were obtained by means of regression analysis in the plot of failure stress versus log of time to failure.

(5)

where a is the crack depth and b is its half width. For this analysis, only three stress rates were considered: 0.01, 1 and 100 MPa/s (N = 4), since the objective was to observe the occurrence of slow crack growth by comparing critical crack sizes at different stress rates. Specimens selected for this analysis were those with strength as close as possible to the calculated mean value of their groups.

2.3.

(stress rate of 1 MPa/s) in each one of the pHs tested. One Vickers indentation (load of 4.9 N and dwell time of 20 s) was made on the polished surface of each specimen in air (∼22 ◦ C, ∼60% RH), using a hardness machine (Mitutoyo MVK-H-3, Tokyo, Japan). Measurement of crack length was made in an optical microscope within 30 s after indentation. Fracture toughness was calculated according to the following equation developed by Anstis et al. [23]:

Hardness and fracture toughness

Hardness and fracture toughness were determined using 10 polished discs stored in ambient air, and 10 fragments of discs of each porcelain recovered from the dynamic fatigue test

3.

Results

Table 1 shows the mean fracture stresses of both porcelains according to the pH and stress rate. The strength values increased by 33% with the increase in the stress rate from 0.01 to 10 MPa/s, regardless of the material and pH. The mean failure stresses measured at 100 MPa/s were lower than those obtained at 10 MPa/s, except for porcelain VM7 stored in pH 7.0, and porcelain UD in pH 3.5. The strength values obtained for porcelain VM7 were approximately 34% higher than those of porcelain UD, regardless of the pH and stress rate (Fig. 1); however, these differences were not always statistically significant (Table 1).

817

d e n t a l m a t e r i a l s 2 4 ( 2 0 0 8 ) 814–823

Table 1 – Mean values (and standard deviations) of failure stress (MPa) as a function of the stress rate, and the pH of the storage medium for both porcelains studied pH

3.5

7.0

10.0

Stress rate (MPa/s)

Porcelain VM7

UD

100 10 1 0.1 0.01

80.7 (13.4) a,b 87.2 (14.3) a 75.9 (17.3) a,b 65.9 (14.7) b 64.7 (11.5) b,c

66.3 (6.4) b 66.0 (7.4) b 56.9 (9.2) b,c 55.3 (5.5) b,c 47.7 (6.8) c

100 10 1 0.1 0.01

90.7 (16.9) a 89.8 (14.3) a 80.8 (15.0) a,b 75.1 (10.0) a,b 68.4 (7.9) b

63.9 (9.5) b 65.6 (8.5) b 57.2 (7.4) b,c 56.1 (8.0) b,c 52.4 (7.1) b,c

100 10 1 0.1 0.01

78.5 (15.7) a,b 87.5 (13.6) a 74.4 (13.9) a,b 64.1 (11.3) b,c 62.8 (8.7) b,c

57.2 (6.7) b,c 60.0 (7.6) b,c 53.4 (8.5)b,c 52.2 (5.5) b,c 45.9 (2.8) c

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

The dynamic fatigue parameters (n and  0 , Eq. (4)) are shown in Table 2 for both porcelains as a function of the pH. For porcelain VM7, the values of n (SCG susceptibility coefficient) obtained in the different pHs were similar and varied from 29.9 to 31.2. The  0 value (scaling parameter) obtained in pH 7.0 for VM7, was higher than that obtained in the other pHs, which were similar. For porcelain UD, the n values obtained in pHs 7.0 and 10.0 were similar (40.8 and 39.6, respectively), and higher than that obtained in pH 3.5 (26.5). With respect to the scaling parameter ( 0 ), the value obtained for porcelain UD in pH 10.0 was lower than those obtained in pHs 3.5 and 7.0, which were similar (Table 2). The comparison of the n values of both materials shows that in pHs 7.0 and 10.0 porcelain UD obtained the highest values, however in pH 3.5 porcelain VM7 obtained the highest value. When the  0 of both materials are compared, it is possible to note that porcelain VM7 obtained higher values compared to UD in all pHs. Fig. 2 represents the lifetime curves for both porcelains in the three studied pHs. The slopes of the curves in Fig. 2 are

Table 2 – Fatigue parameters (standard deviation) for both porcelains studied according to the pH of the storage medium (n: stress corrosion coefficient;  0 : scaling parameter;  10 : predicted strength after 10 years of storage) Porcelain

pH

n

 0 (MPa)

 10 (MPa)

UD

3.5 7.0 10.0

26.5 (4.2) 40.8 (10.2) 39.6 (8.9)

57.81 (0.02) 58.33 (0.02) 53.23 (0.02)

32.9 41.1 36.8

VM7

3.5 7.0 10.0

29.9 (8.1) 31.2 (7.1) 30.4 (7.6)

73.12 (0.03) 79.65 (0.02) 71.80 (0.03)

46.8 51.0 46.0

directly related to the materials’ n value (the higher the n value, the lower the slope). Thus, the lower n value of porcelain UD compared to VM7 in pH 3.5 results in the higher slope observed for the first in Fig. 2a. The fracture stress correspondent to 10 years ( 10 ) was calculated for each material using the equation obtained in the regression analysis (these values are also presented in Table 2). Mean critical crack sizes measured in the fractographic analysis are shown in Fig. 3. Crack sizes were higher in porcelain UD, regardless of the pH and stress rate. Also, it is possible to note a decrease in crack size with the increase in stress rate for both porcelains in all pHs. For porcelain VM7, crack sizes in pH 7.0 were slightly smaller compared to the other pHs, however the values of c had similar decreasing rate with the increase in stress rate for the three investigated pHs. Crack sizes measured in the fracture surface of porcelain UD did not vary significantly with the pH (larger sizes, however, were usually observed in pH 3.5 and 10.0 compared to pH 7.0), but it is possible to note in Fig. 3 that, compared to VM7, the c values had higher decreasing rate with the increase in stress rate for all pHs. Examples of fracture surfaces and fracture origins of both porcelains are shown in Fig. 4, where it is possible to note that porcelain VM7 has a relatively smoother fracture surface compared to UD. All specimens analyzed showed semi-elliptical surface cracks as the fracture origin. Table 3 shows the mean values of fracture toughness (KIc ) and hardness for all experimental conditions. The values of KIc of both porcelains did not vary significantly according to the

Table 3 – Fracture toughness (KIc ) and Vickers hardness (standard deviation) as a function of storage condition for both porcelains tested

Fig. 1 – Fracture stress as a function of stress rate for both porcelains in the different pHs studied.

Storage condition

UD

VM7

KIc (MPa m1/2 )

Air pH 3.5 pH 7.0 pH 10.0

1.07 (0.07) a 0.97 (0.09) a 1.00 (0.12) a 0.96 (0.11) a

0.81 (0.04) b 0.78 (0.05) b 0.80 (0.10) b 0.77 (0.02) b

Hardness (GPa)

Air pH 3.5 pH 7.0 pH 10.0

7.2 (0.5) a 6.8 (0.6) a,b 6.5 (0.5) a,b 6.2 (0.6) b,c

6.5 (0.7) a,b 6.3 (0.5) b,c 6.2 (0.6) b,c 5.7 (0.4) c

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

818

d e n t a l m a t e r i a l s 2 4 ( 2 0 0 8 ) 814–823

Fig. 3 – Crack size as a function of stress rate for both porcelains in the different pHs studied.

Table 4 – Semi-quantitative chemical analysis by X-ray fluorescence (in mass%) for porcelains studied

Fig. 2 – Plots of the time to failure for both porcelains stored in a solution with pH 3.5 (a), pH 7.0 (b), and pH 10 (c). The independent variable is the log of time in seconds and the dependent variable is the log of failure stress. The time axis is labeled for 1 day (1d), 1 year (1y) and 10 years (10y). Data points were omitted for the sake of clarity.

storage condition; however, the fracture toughness of porcelain VM7 (∼0.8 MPa m1/2 ) was always significantly lower than that of porcelain UD (∼1.0 MPa m1/2 ). The hardness of both materials decreased with the increase in the pH, but only the difference between the group stored in air and the group stored in pH 10.0 was statistically significant. The hardness of porcelain UD was higher than that of VM7 regardless of the storage condition, however, these differences were not statistically significant. XRD analysis of the raw powder of porcelain UD detected a large band (representative of the amorphous phase) and

Component

VM7

UD

SiO2 Al2 O3 K2 O Na2 O CaO ZrO2 MgO

66.8 15.6 10.5 3.2 2.7 0.8 0.0

60.8 16.4 13.0 5.1 2.9 0.4 0.5

diffraction peaks corresponding to the crystalline leucite phase (KAlSi2 O6 ). The XRD pattern of porcelain VM7 presented only a large band corresponding to the amorphous phase. SEM analysis (Fig. 5a) showed the presence of second-phase particles (leucite, according to the XRD analysis) in porcelain UD. The leucite content was 13.2 vol% (mean particle size of 1.4 ␮m) and the distribution of the particles in the glassy matrix was heterogeneous (particles were grouped in clusters, Fig. 5b). It was also possible to observe in porcelain UD the presence of some cracks around the leucite particles and clusters. Some acicular particles were also observed in the surface of porcelain UD (Fig. 5a). These particles are most likely precipitates formed after the reaction of the porcelain with the hydrofluoric acid. Porcelain VM7 (Fig. 5c) did not present second-phase particles or cracks after etching with HF; however, it was possible to note regions with different corrosion rates, probably related to the starting glass powder. Table 4 shows the chemical composition (in mass%) of all porcelains studied. For both materials the main components were SiO2 , Al2 O3 , K2 O, Na2 O and CaO. Porcelain VM7 presented higher amount of SiO2 (66.8%) than UD (60.8%). With respect to K2 O and Na2 O, porcelain UD presented higher amounts compared to VM7. The amounts of Al2 O3 and CaO were similar for both materials.

4.

Discussion

The present study showed that the effect of pH on the susceptibility to slow crack growth depends on the porcelain studied. The values of stress corrosion coefficient obtained for porcelain VM7 were around 30 (Table 2) and indicated that pre-

d e n t a l m a t e r i a l s 2 4 ( 2 0 0 8 ) 814–823

819

Fig. 4 – SEM images of fractured surfaces of the materials tested in biaxial flexural strength test: (a) and (b) are surfaces of porcelain VM7 tested in pH 7.0 with stress rates of 100 and 0.01 MPa/s, respectively; (c) and (d) are surfaces of porcelain UD tested in pH 3.5 with stress rates of 100 and 0.01 MPa/s, respectively. The arrows indicate the crack front, which delimited the regions of slow and fast crack growth (SCG and FCG, respectively). The fracture origins were confirmed by OM. The location of initial flaw is indicated as “origin” at the specimen surface.

existing flaws in this material will have similar growth rates over time regardless of the saliva’s pH. Conversely, the leucitebased porcelain UD was more susceptible to slow crack growth in lower pH, given that its n value was the lowest of the whole experiment when the pH was 3.5. Results from fractographic analysis support these findings, since the crack growth rates of VM7 were similar in all pHs, while UD showed the highest rate in acid pH (Fig. 3). The higher susceptibility of UD to slow crack growth in acid pH is detrimental for this material providing the pH of the oral cavity becomes acid every time an individual consumes carbohydrates or acidic food [14,24]. The lower stress corrosion coefficient obtained for porcelain UD in acid pH was somewhat unexpected, as some works indicated that this pH would result in decreased susceptibility to slow crack growth [16,17]. The lower n value obtained by porcelain UD may be explained by findings of another study, which showed increased ion dissolution (mainly K, Na, Si, and Al) when one porcelain was immersed in acetic acid (pH 2.4) compared to neutral pH solution [25]. Despite the unfavorable behavior of porcelain UD in acid pH, it is important to consider that in pHs 7.0 and 10.0, its n values were higher than those of

porcelain VM7, which could indicate lower crack growth rates in such environments. However, this finding was not corroborated by the fractographic analysis, since crack growth rates of porcelain UD were higher than those of porcelain VM7 in pHs 7.0 and 10.0 (Fig. 3). Also, these higher n values of UD compared to VM7 cannot be explained by the fact that glasses with higher amounts of alkaline ions (Na and K) are more susceptible to SCG [26], since the sum of these oxides was higher for UD (Table 4). Even considering that some of these oxides were consumed to form leucite particles (KAlSi2 O6 ) in porcelain UD, its glassy matrix contains higher amounts of alkaline oxides (15.7 and 11.0 mol% for UD and VM7, respectively). The explanation for the higher n values of UD in pHs 7.0 and 10.0 compared to VM7 may reside in the possibility of this material presenting the so-called R-curve behavior (crack resistance curve), i.e. an increase in fracture toughness with the increase in flaw size. It has been demonstrated that leucite-based dental porcelains present such behavior as a consequence of friction between rough crack surfaces caused by crack deflection around leucite agglomerates [27]. Therefore, the mechanical grip between rough surfaces in the crack wake may have shielded the crack tip, leading to increased

820

d e n t a l m a t e r i a l s 2 4 ( 2 0 0 8 ) 814–823

Fig. 6 – Fracture stress as a function of the square root of crack size (c−1/2 ) for both porcelains.

resistance to slow crack propagation and consequently higher n values in pHs 7.0 and 10.0. The better behavior of porcelain VM7 compared to UD in terms of early strength ( 0 , Table 2) may be explained by the presence of smaller and less numerous microstructural defects in the first [28]. In fact, fractographic analysis (Figs. 3 and 4) showed that VM7 had significantly smaller cracks compared to UD, and Fig. 6 shows a positive correlation between fracture stress ( f ) and c−1/2 values for both porcelains, as predicted by the Griffith–Irwin equation [29]: f =

Fig. 5 – Micrographs of the polished surfaces of porcelains UD (a) and VM7 (c) etched with 2% hydrofluoric acid (HF) for 15 s. The image obtained in the image analyzer software (c) shows the leucite particles delimited manually.

KIc Yc1/2

(7)

The larger cracks observed in porcelain UD are related to the presence of leucite particles in this material. Although leucite particles lead to increased fracture toughness by means of crack deflection [30,31], it is important to note that such particles also cause spontaneous crack formation, mainly when grouped in clusters, resulting in strength degradation [28]. In fact, it was observed that the larger cracks in porcelain UD were usually related to large leucite agglomerates (Fig. 4c). Crack size is not the only variable affecting  f as indicated by Eq. (7). The crack shape factor, Y, also plays a role and depends on the aspect ratio of the semi-elliptical crack, b/a, where a and b are the depth and the half width of superficial cracks, respectively. The b/a values of porcelain UD were almost constant around 1.5, irrespective of pH and stress rate (Fig. 7). For porcelain VM7, the values of b/a in pH 3.5 and 7.0 were around 2.2, regardless of stress rate, showing that cracks were more elongated in this material than in porcelain UD. In pH 7.0, however, the value of b/a increased with the decrease in stress rate from ∼1.1 at 102 MPa/s to ∼2.2 at 10−2 MPa/s, showing that the crack evolved from an almost semicircular shape to an elongated semi-elliptical shape. This trend is usually observed and is related to the point at the crack front in which the Y value is maximum: for semicircular flaws, it is located at the outer surface of the sample, Ys , which favors initial rapid lateral growth of the crack, and for semi-elliptical flaw, it is located at the deepest point of the crack, Yd [32]. This change in crack shape, however, does not seem to have

d e n t a l m a t e r i a l s 2 4 ( 2 0 0 8 ) 814–823

Fig. 7 – Crack aspect ratio (b/a) as a function of stress rate for both porcelains in the different pHs studied (a is crack depth and b is crack width).

affected the slow crack growth phenomenon, as the n value of VM7 did not vary as a function of pH, as stated earlier. This finding is in accordance to a previous work [33], in which the change in precrack shape in a soda-lime silica glass had little influence on SCG parameters determined by dynamic fatigue method. The Newman and Raju equations [34] for the stress intensity shape factor were used to calculate Ys and Yd considering the measured crack dimensions for each evaluated specimen and the highest value was considered as the actual Y value. The average value of Y determined for VM7 was 1.36, which was higher than that of porcelain UD (1.20), because of the more elongated crack shape in the first (Fig. 7). Using the calculated Y value, the measured fracture stress and crack size, the material’s fracture resistance (KR ) was determined from an equation similar to Eq. (7) and the results are plotted as a function of crack size in Fig. 8. As expected, porcelain VM7 presented constant KR values with increasing crack size (flat R-curves) in all pHs, since its microstructure was composed only by a glassy phase without reinforcing particles (Fig. 5c), which resulted in smooth fracture surfaces (Fig. 4a and b). The average KR value for VM7 was 1.10 MPa m1/2 , which was significantly higher than the KIc value determined by IF method (0.81 MPa m1/2 , Table 3). The reason for this difference may be the high crack aspect ratio of this porcelain, since Eq. (6) was derived for a penny-like radial/median cracks [23].

Fig. 8 – Fracture toughness (KR ) as a function of crack size (c) for both porcelains in the different pHs studied.

821

The KR value of porcelain UD in pH 7.0 fractured at the higher stress rate (shortest crack size) was 1.10 MPa m1/2 (Fig. 8), which was close to the value determined by IF method (1.07 MPa m1/2 , Table 3). With the increase in crack size, KR tended to increase in all three pHs, but with different increments: 5, 13 and 16% in pH 3.5, 10.0 and 7.0, respectively, with crack size ranging from ∼200 to 500 ␮m (Fig. 8). These results confirmed that porcelain UD showed rising R-curve behavior and supported the hypothesis that the high SCG coefficients, n, in pH 7.0 and 10.0 (∼40, Table 2) were related to crack deflection around leucite agglomerates. Lifetime curves presented in Fig. 2 help to better understand differences observed in fatigue parameters of both porcelains [35], since curves with lower slopes indicate lower strength degradation over time [36]. Hence, the unfavorable behavior of porcelain UD in acid pH can be noted by the higher slope of its curve compared to VM7 in Fig. 2a. Moreover, it is observed in Table 2 that UD’s predicted strength after 10 years ( 10 ) in pH 3.5 is 43% lower than its early strength ( 0 ). As for VM7, the  10 value in pH 3.5 was 36% lower than its  0 . Direct comparison of  10 values of both materials shows that predicted strength after 10 years in pH 3.5 is 27% higher for VM7. The comparison of lifetime curves of both porcelains in pHs 7.0 and 10.0 also indicates that VM7 has a better overall mechanical behavior than UD, since VM7s predicted strength after 10 years is higher than that of UD. Though basic pH did not alter the n values of materials compared to neutral pH, it is important to note that for both porcelains, comparison of  0 values obtained in these two pHs (Table 2) showed that values obtained in pH 7.0 were about 9% higher than those obtained in pH 10.0. These observations may be explained by the fact that specimens were stored for 10 days in each experimental solution before the strength test was carried out. Such strength degradation may be related to the dissolution of glass surfaces promoted by alkaline solution containing NaOH [18]. The surface corrosion seems to have lowered the material’s fracture resistance (KR ) in basic saliva in both porcelains (Fig. 8), lowering the  0 value. The change in early strength without a subsequent alteration in the stress corrosion coefficient indicates that the kinetics of the reaction between ions OH+ and SiO4 4− is an important factor to comprehend the slow crack growth phenomenon in the porcelains studied. It appears that such reaction at the crack tip occurs at high velocity in the early times (i.e. the first 10 days of storage), causing the observed decrease in strength values with the increase in pH. After this storage period, it is likely that the reaction velocity decreases significantly because of saturation of the solution, or because of the lack of agitation. Thus, when the dynamic fatigue test was carried out after 10 days of storage, the low velocity of the reaction in basic pH led to similar n values compared to neutral pH. A decrease in  0 was also noted for porcelain VM7 in pH 3.5, compared to pH 7.0, without a change in the n value (Table 2). In this case, the strength degradation may be related to the rapid dissolution of ions in the first days of storage as a consequence of the exchange of H+ ions from the saliva for alkaline ions (Na+ and K+ ) from the porcelain. This process tends to cease with time because the concentration of alkaline ions increases rapidly at the crack tip, elevating the pH to 12 in this region [16].

822

d e n t a l m a t e r i a l s 2 4 ( 2 0 0 8 ) 814–823

Care should be taken when using the results of the present study to predict clinical behavior of tested porcelains because the oral environment presents some important differences from the storage media used here. While in this study solution had a constant pH, the oral pH is known to suffer frequent alterations according to the production of acids by the dental plaque and according to the pH of the food ingested. Moreover, other features of the oral cavity were not addressed here, such as application of cyclic stresses during the storage period, temperature variations, agitation of saliva, brushing of surfaces, and presence of biofilm. It is also important to consider that the results of this study are valid for clinical situations in which the fracture initiating flaws are in direct contact with saliva, as in the case of fixed partial denture connectors. For porcelains restorations like inlays, onlays and veneers, the results may not be completely valid, because clinical studies have demonstrated that in such restorations fracture usually starts from defects present in the cementing face [3]. In such cases, the environment characteristics are influenced by the luting cement in contact with the porcelain surface.

[7]

[8] [9] [10]

[11]

[12] [13]

[14]

[15]

5.

Conclusion

Based on the results of this study, it was possible to conclude that the effect of pH on the stress corrosion susceptibility (n) depended on the porcelain studied. While the n value of the vitreous porcelain (VM7) was not affected by the pH, the leucite-based porcelain (UD) presented lower n value in acid pH compared to basic and neutral. For both porcelains, storage in acid or basic pH resulted in strength degradation. The overall mechanical behavior of VM7 was considered better than UD because the results of lifetime prediction showed that the predicted strength after 10 years was always higher for the first regardless of the pH studied.

[16] [17]

[18]

[19]

[20] [21]

Acknowledgments The authors acknowledge the Brazilian agencies FAPESP and CNPq for the financial support of the present research.

[22]

[23]

references [24] [1] Haselton DR, Diaz-Arnold AM, Hillis SL. Clinical assessment of high-strength all-ceramic crowns. J Prosthet Dent 2000;83:396–401. [2] Kelly JR, Giordano R, Pober R, Cima MJ. Fracture surface analysis of dental ceramics: clinically failed restorations. Int J Prosthodont 1990;3:430–40. [3] Thompson JY, Anusavice KJ, Naman A, Morris HF. Fracture surface characterization of clinically failed all-ceramic crowns. J Dent Res 1994;73:1824–32. [4] Myers ML, Ergle JW, Fairhurst CW, Ringle RD. Fatigue failure parameters of Ips-empress porcelain. J Dent Res 1994;73:136–136. [5] Ritter JE. Predicting lifetimes of materials and material structures. Dent Mater 1995;11:142–6. [6] De Aza AH, Chevalier J, Fantozzi G, Schehl M, Torrecillas R. Crack growth resistance of alumina, zirconia and zirconia

[25]

[26]

[27]

[28]

toughened alumina ceramics for joint prostheses. Biomaterials 2002;23:937–45. Wiederhorn SM. Influence of water vapor on crack propagation in soda-lime glass. J Am Ceram Soc 1967;50:407–14. Wan KT, Lathabai S, Lawn BR. Crack velocity functions and thresholds in brittle solids. J Eur Ceram Soc 1990;6:259–68. Michalske TA, Bunker BC. Slow fracture model based on strained silicate structures. J Appl Phys 1984;56:2686–93. Morena R, Beaudreau GM, Lockwood PE, Evans AL, Fairhurst CW. Fatigue of dental ceramics in a simulated oral environment. J Dent Res 1986;65:993–7. Guazzato M, Albakry M, Ringer SP, Swain MV. Strength, fracture toughness and microstructure of a selection of all-ceramic materials. Part I. Pressable and alumina glass-infiltrated ceramics. Dent Mater 2004;20:441–8. Featherstone JD. Modeling the caries-inhibitory effects of dental materials. Dent Mater 1996;12:194–7. Itthagarun A, Wei SH, Wefel JS. Morphology of initial lesions of enamel treated with different commercial dentifrices using a pH cycling model: scanning electron microscopy observations. Int Dent J 1999;49:352–60. Ten Cate JM, Duijsters PP. Alternating demineralization and remineralization of artificial enamel lesions. Caries Res 1982;16:201–10. Thylstrup A, Fejerskov O. Textbook of clinical cariology. Copenhagen: Munksgaard; 2001. Wiederhorn SM, Johnson H. Effect of eletrolyte pH on propagation in glass. J Am Ceram Soc 1973;56:192–7. Ritter J, Laporte R. Effect of test environment on stress-corrosion susceptibility of glass. J Am Ceram Soc 1975;58:265–7. Tomozawa M, Oka Y, Wahl MJ. Glass surface cracks caused by alkaline solution containing an alkaline-earth element. J Am Ceram Soc 1981;64:32–3. Simmons C, Freiman S. Effect of corrosion processes on subcritical crack-growth in glass. J Eur Ceram Soc 1981;64:683–6. ASTM-C1161. Flexural strength of advanced ceramics at ambient temperature. 2002. ASTM-C1368-00. Standard test method for determination of slow crack growth parameters of advanced ceramics by constant stress-rate flexural testing at ambient temperature. Annual book of ASTM, vol. 15. American Society for Testing and Materials; 2001. p. 1–9. Morena R, Lockwood PE, Fairhurst CW. Fracture toughness of commercial dental porcelains. Dent Mater 1986;2: 58–62. Anstis G. A critical of indentation techniques for measuring fracture toughness: in direct crack measurements. J Am Ceram Soc 1981;64:533–8. Featherstone JD, Holmen L, Thylstrup A, Fredebo L, Shariati M. Chemical and histological changes during development of artificial caries. Caries Res 1985;19:1–10. Milleding P, Haraldsson C, Karlsson S. Ion leaching from dental ceramics during static in vitro corrosion testing. J Biomed Mater Res 2002;61:541–50. Sglavo VM, Bertoldi M. Vickers indentation: a powerful tool for the analysis of fatigue behavior on glass. In: Indentation techniques in ceramic material characterization. Westerville: The American Ceramic Society; 2004. Yoshimura HN, Cesar PF, Miranda Jr WG, Gonzaga CC, Okada CY. Fracture toughness of dental porcelains evaluated by IF, SCF, and SEPB methods. J Am Ceram Soc 2005;88:1680–3. Cesar PF, Yoshimura HN, Miranda Jr WG, Miyazaki CL, Muta LM, Rodrigues Filho LE. Relationship between fracture toughness and flexural strength in dental porcelains. J Biomed Mater Res B: Appl Biomater 2006;78:265–73.

d e n t a l m a t e r i a l s 2 4 ( 2 0 0 8 ) 814–823

[29] Yoshimura HN, Molisani AL, Narita NE, Manholeti JLA, Cavenaghi JM. Mechanical properties and microstructure of zinc oxide varistor ceramics. Mater Sci Forum 2006;530:408–13. [30] Mackert Jr JR, Twiggs SW, Russell CM, Williams AL. Evidence of a critical leucite particle size for microcracking in dental porcelains. J Dent Res 2001;80:1574–9. [31] Cesar PF, Yoshimura HN, Miranda Jr WG, Okada CY. Correlation between fracture toughness and leucite content in dental porcelains. J Dent 2005;33:721–9. [32] Quinn GD. Fractography of ceramics and glasses. Washington: National Institute of Standards and Technology; 2007.

823

[33] Sglavo V, Green D. Influence of indentation crack configuration on strength and fatigue behavior of soda-lime silicate glass. Acta Metall Mater 1995;43:965–72. [34] ASTM-1421. Standard test methods for determination of fracture toughness of advanced ceramics at ambient temperature. 1999. [35] Myers ML, Ergle JW, Fairhurst CW, Ringle RD. Fatigue characteristics of a high-strength porcelain. Int J Prosthodont 1994;7:253–7. [36] Twiggs SW, Fairhurst CW, Lockwood PE, Ringle RD. Cyclic fatigue of a model feldspathic porcelain. Dent Mater 1995;11:273–6.