journal of the mechanical behavior of biomedical materials 47 (2015) 1 –11
Available online at www.sciencedirect.com
www.elsevier.com/locate/jmbbm
Research Paper
Fracture toughness of esthetic dental coating systems by nanoindentation and FIB sectional analysis Christina Martina Pecnik, Diana Courty, Daniel Muff, Ralph Spolenakn Laboratory for Nanometallurgy, Department of Materials, ETH Zurich, Zurich, Switzerland
art i cle i nfo
ab st rac t
Article history:
Improving the esthetics of Ti-based dental implants is the last challenge remaining in the
Received 10 November 2014
optimization process. The optical issues were recently solved by the application of highly
Received in revised form
and selectively reflective coatings on Ti implants. This work focuses on the mechanical
28 February 2015
durability of these esthetic ceramic based coating systems (with and without adhesion
Accepted 10 March 2015
layers).
Available online 19 March 2015
The coating systems (Ti–ZrO2, Ti–Al–ZrO2, Ti–Ti–Al–ZrO2, Ti–Ag–ZrO2, Ti–Ti–Ag–ZrO2,
Keywords:
Ti–Bragg and Ti–TiO2–Bragg) were subjected to nanoindentation experiments and exam-
Fracture toughness
ined using scanning electron microscopy and focused ion beam cross sectional analysis.
Dental implants
Three coating systems contained adhesion layers (10 nm of Ti or 60 nm of TiO2 layers). The
Thin films
fracture toughness of selected samples was assessed applying two different models from
Nanoindentation
literature, a classical for bulk materials and an energy-based model, which was further
FIB
developed and adjusted.
Titanium Ceramics
The ZrO2 based coating systems (total film thicknesso200 nm) followed a circumferential cracking behavior in contrast to Bragg coated samples (total film thickness around 1.5 μm), which showed radial cracking emanating from the indent corners. For Ti–ZrO2 samples, a fracture toughness between 2.70 and 3.70 MPa m1/2 was calculated using an energy-based model. The classical model was applied to Bragg coated samples and their fracture toughness ranged between 0.70 and 0.80 MPa m1/2. Furthermore, coating systems containing an additional layer (Ti–Ti–Al–ZrO2, Ti–Ti–Ag–ZrO2 and Ti–TiO2-Bragg) showed an improved adhesion between the substrate and the coating. The addition of a Ti or TiO2 layer improved the adhesion between substrate and coating. The validity of the models for the assessment of the fracture toughness depended on the layer structure and fracture profile of the samples investigated here (classical model for thick coatings and energy-based model for thin coatings). & 2015 Elsevier Ltd. All rights reserved.
n Correspondence to: Laboratory for Nanometallurgy, Department of Materials, ETH Zurich, Vladimir-Prelog-Weg 5, 8093 Zurich, Switzerland. Tel.: þ41 44 632 25 90; fax: þ41 44 632 11 01. E-mail address: [email protected] (C.M. Pecnik).
http://dx.doi.org/10.1016/j.jmbbm.2015.03.006 1751-6161/& 2015 Elsevier Ltd. All rights reserved.
2
1.
journal of the mechanical behavior of biomedical materials 47 (2015) 1 –11
Introduction
Dental implants have revolutionized the reconstructive dentistry since the late 1970’s and offer a reliable treatment to reestablish a healthy and functional dentition after tooth-loss (Brånemark et al., 1977). Titanium (Ti) is still considered as the ‘gold-standard’ for dental reconstructions, since it exhibits a high fatigue and corrosion resistance, biocompatibility and excellent osseointegration as shown in long-term studies (Albrektsson et al., 1981; Andersson, 1995; Buser et al., 2012; Dierens et al., 2012). An important aspect, though, is the esthetical appearance of the peri-implant soft tissue: metallic reconstructions can negatively influence the esthetics caused by recession and/or dark discolorations of the gingival tissue and, therefore, impair the patient's approval of the treatment’s results (Joss-Vassalli et al., 2010; Jung et al., 2007; van Brakel et al., 2011). This disadvantage of the Ti implant can be remedied by ceramic based coatings (Pecnik et al., 2014a, 2014b), which showed significantly improved optical properties compared to Ti and were suggested as esthetic dental coatings for future application. In addition to the optical analysis, a mechanical characterization of these thin film systems is needed since their performance should be reliable during service. One simple way to assess the mechanical properties of thin films is by nanoindentation (Doerner and Nix, 1986; Oliver and Pharr, 1992). For thin films, the hardness still can be determined at small indentation depths (o10% of the film thickness) (Saha and Nix, 2002), whereas the assessment of the fracture toughness becomes a challenge. Various models on fracture mechanisms in thin films have been described in literature; however, the use of the appropriate method depends on the thickness, cracking scenario and shape of the load–displacement curve of the tested film (Bull, 2011; Chen and Bull, 2007, 2011; Schiffmann, 2011). Despite the fact that sharp indentation is infrequently occurring in the periimplant region respectively in the soft tissue part of the implant, nanoindentation is a convenient technique to measure not only the material’s properties, but it gives also information about the adhesion and/or cohesion of coatings (Chen and Bull, 2011; Gerberich and Cordill, 2006; Lu et al., 2013). This method allows to characterize the fracture toughness of thin films under clearly defined loading conditions, contrary to the complex actual loading scenarios present in the oral environment The fracture toughness of a material is not only a property that describes the resistance of further cracking, but it is also helpful to value its reliability under mechanical stress. Nanoindentation with a cube corner tip was conducted at different loads to create cracking in the films. Furthermore, focused ion beam (FIB) cross sections of the samples were analyzed in order to reveal the extent of fracture and deformation in the film, substrate and in between them after indentation. Some of the investigated coating systems exhibited very thin film thicknesses (below 200 nm), which is a challenging task for the implementation of nanoindentation experiments and their analysis. Accordingly, different models such as the classical model used for bulk materials (Lawn et al., 1980) and energy-based models by Bull and Chen (2011), Chen and Bull (2007) were applied to calculate the fracture toughness of the coating systems. Furthermore, selected
coating systems contained adhesion layers, which ultimately provided information about the interplay between coating and substrate visible through FIB cross sections. Moreover, the purpose of these experiments was also to compare the results qualitatively between the coating systems and finally, discuss the validity of the quantitative results and calculations from the applied models.
2. Overview of models for the determination of fracture toughness by nanoindentation The determination of the fracture toughness of very thin films (o500 nm) is a difficult task for which, in the last 30 years, different models and techniques were developed. In general, the fracture toughness for mode I can be described by the following equation: pffiffiffiffiffiffiffiffi ð1Þ KIc ¼ EGc where KIc is also called the critical stress-intensity factor, E is the Young modulus and Gc is the critical strain energy release rate, which corresponds to the dissipated energy of a crack per unit area. Mode I describes the tensile loading acting perpendicularly on a crack (Rösler et al., 2007). After indentation in a brittle material, there are different kinds of crack initiation patterns and stress fields. Most applied crack geometries are halfpenny-shaped and Palmquist (or radial) cracks for brittle materials, which were classified by Cook and Pharr (1990). It is, however, difficult to distinguish between crack geometries from the top view of an indent. Mainly, the analysis of the half-penny crack system is used to determine the toughness of brittle materials (Bull, 2011). On the assumption of the half-penny crack system Lawns et al. developed an equation to determine the fracture toughness after indentation by measuring the length of cracks that originate at the corners of the indent (Lawn et al., 1980): 1=2 E P 3=2 ð2Þ Kc ¼ α H c where Kc is now addressed to the fracture toughness since the loading condition in indentation experiments does not correspond to a pure mode I loading. H is the hardness, P is the applied load, and c is the length of the crack emanating from the corners of the indent. The empirical constant α depends on the indenter geometry and crack system, but there is a range of different values in literature for each tip geometry and crack pattern, which will be discussed later on. This equation, however, can only be applied for bulk brittle materials or sufficiently thick brittle films, as suggested by Bull (2011), Chen (2012). In contrast to the conventional, classical indentation methods, which are more suitable for bulk materials, nanoindentation is a well-established method to determine the Young modulus and hardness of thin films. For fracture toughness experiments, one has to facilitate the creation of cracks at the corners of the indent. Suitably, the cube corner tip exhibits sharp sides that induce high enough stress intensities and, therefore, even lowers the threshold of load for cracking (Harding et al., 1995; Pharr, 1998). As the coating thickness decreases, further problems could arise for the assessment of the fracture toughness: radial cracking may no longer be observed and with increasing load,
journal of the mechanical behavior of biomedical materials 47 (2015) 1–11
the effect from the substrate becomes more significant. Alternative methods have been developed by Bull (2011), Chen and Bull (2006, 2007, 2011), Chen (2012), where energy-based models were applied for the determination of the fracture toughness for very thin films. In their alternative approach the total work of indentation (Wtot) corresponded to the area below the load– displacement (P–h) curve, which can be further split in different parts of the total work such as elastic, plastic and fracture part. Depending on the presence of features in the P–h curve, two different methods are described to determine the fracture toughness. Here, the model without features (e.g. no slope change or discontinuities) in the P–h curve will be described in detail.
2.1.
Energy-based model for P–h curves without features
The total work of indentation can be defined as (Bull, 2011; Chen and Bull, 2007) Wtot ¼ Welast þ Wplast þ Ufrac þ Wother
ð3Þ
where Welast is the work of elastic deformation, Wplast is the work of plastic deformation, Ufrac is the dissipated energy during fracture, and Wother is the work due to heat dissipation and/or creep during indentation. Wother will be neglected in the following since it is assumed to be a small term contributing to the total work. This then reduces Eq. (3) to the following equation: Wtot ¼ Welast þ Wplast þ Ufrac
ð4Þ
Another energy term can be defined, which consists of the plastic and fracture energy: Uirr ¼ Wplast þ Ufrac
ð5Þ
where Uirr is the irreversible energy, whereas the reversible
3
plastic behavior is not considered in this case. After an indentation experiment, Wtot, Welast, Uirr can directly be obtained from the resulting load displacement curve. However, the plastic work Wplast has to be determined using a different approach (Cheng et al., 2002; Lawn and Howes, 1981): 2 3 h
2
h
f f 61 3 hm þ 2 hm Wplast ¼ 1 6 2 4 Wtot h 1 hmf
3
7 7 5
ð6Þ
The obtained fracture energy Ufrac can then be further used to define the critical strain energy release rate Gc: Gc ¼
Ufrac Afrac
ð7Þ
where Afrac is the area produced during cracking. Both interfacial and coating cracking can be considered in the total fracture area Afrac. Finally, the fracture toughness can be calculated using Eq. (1) with the substitution of KIc with Kc and E with Er: pffiffiffiffiffiffiffiffiffiffi ð8Þ Kc ¼ Er Gc
3.
Materials and methods
3.1.
Coating systems
The different coating systems in this study are illustrated in Fig. 1 and is based on the design strategy described in Pecnik et al. (2014a, 2014b). Additionally, some coating systems were deposited with an adhesion layer (Ti or TiO2). The deposition parameters are displayed in Table 1.
Fig. 1 – Layer structure of the coating systems. Seven coating systems were characterized in this study: all were deposited at a process chamber pressure of 2 mTorr. Additionally, the three samples contained adhesion layers (10 nm of Ti for ZrO2 based coating systems with Al resp. Ag and 60 nm of TiO2 for the Bragg coating).
4
journal of the mechanical behavior of biomedical materials 47 (2015) 1 –11
3.2.
Substrate preparation and thin film deposition
Round Ti platelets (Grade 4, ThyssenKrupp) were used as substrate material, and exhibited a thickness of 0.5 mm and a diameter of 15 mm. An automatic grinding and polishing machine (Rotopol-2, Struers) was used for substrate preparation. Initially, the platelets were ground using silicon carbide papers up to 4000 grid. Prior to the polishing step, the platelets were ultrasonically cleaned in ethanol for 5 min. Subsequently, the platelets were polished with a modified 0.02 μm colloidal silica suspension (MasterMet 2, Buehler), containing 10 vol% hydrogen peroxide (Z35%, Sigma-Aldrich). The polished substrates were ultrasonically cleaned with ethanol and acetone (1:1 mixture) for 10 min. The average area roughness of the polished Ti platelets was Sa ¼ 0.005 μm, obtained by atomic force microscopy (Cypher, Asylum Research, USA) (Pecnik et al., 2014a, 2014b). The coating systems (Table 1) were then deposited by reactive magnetron sputtering at room temperature (PVD Products, Inc.). The base pressure in the deposition chamber was 2 mTorr. The thickness of the reactively deposited zirconia (ZrO2) layer was selected for a pink interference color (165.5 nm). Both metallic intermediate layers, aluminum (Al, 99.9%, Sindlhauser Materials GmbH) and silver (Ag, 99.99%, Kurt J. Lesker), exhibited a thickness of 50 nm. Furthermore, an adhesion layer was deposited on three further Ti samples (10 nm of Ti as adhesion layer for ZrO2 based coatings with Al resp. Ag layers, and 10 nm of TiO2 as adhesion layer for the Bragg coating, see Fig. 1).
3.3.
Nanoindentation
The mechanical properties of the substrate and different coating systems were characterized using an Ultra Nanoindentation Tester (UNHT, CSM Instruments by Anton Paar, Switzerland) with a cube corner tip. The Young’s modulus Ei of the indenter tip was 1140 GPa and its Poisson’s ratio νi was 0.07. Displacement-controlled experiments were conducted for 12 different indentation depths (45 to 200 nm in different step sizes). For each indentation depth, the indentation was repeated 20 times with a distance of 10 μm between the indents. Loading and unloading rates were adjusted to the corresponding indentation depth (e.g. depth¼45 nm, loading
resp. unloading rate¼ 450 nm/min) without dwell time keeping the loading rate constant. The results were analyzed using the method by Oliver and Pharr (1992) to determine the hardness H and reduced modulus Er. Load-controlled experiments were conducted for five different loads (10 to 50 mN with 10 mN step size). For each load, the indentation was repeated 20 times with a distance of 30 μm between the indents. Loading and unloading rates were adjusted to the corresponding load (e.g. load¼ 10 mN, loading resp. unloading rate¼ 12 ∙ load [mN/min]¼120 mN/min) without dwell time keeping the loading rate constant.
3.4.
Characterization and evaluation
Indents from load-controlled experiments were analyzed by means of secondary electron imaging (SE) and backscattered electron imaging (BSE) using a Schottky-type field-emission scanning electron microscope (SU 70, Hitachi, Japan). Focused ion beam (FIB, Helios Nanolab 600i, FEI) technique was used to investigate the film cross sections after indentation. A platinum (Pt) layer was deposited on the indent in order to protect the sample surface. After cross-sectioning, the samples were analyzed using the featured Schottky-type fieldemission gun (Helios Nanolab 600i, FEI). The fracture toughness of each sample was evaluated using the previously described models for different loads. Fig. 2 shows the evaluation procedures for both models. Further details to the evaluation procedure will be described in the results section.
4.
Results
4.1.
Indentation results
The mechanical properties of the coating systems and t10he substrate material are shown in Fig. 3. The indentation depths were chosen in a rather small displacement range in order to characterize the properties of each coating itself, neglecting the impact of the substrate. Hardness and reduced modulus values first increased with increasing indentation
Table 1 – Sputter deposition parameters of the coating systems. For Bragg-based coatings, the design wavelength was λ1 ¼480 nm for the first stack (S1) and λ2 ¼590 nm for the second stack (S2). Sample description
Layer thickness [nm]
Ti–ZrO2 Ti–Al–ZrO2
ZrO2: 165.5 ZrO2: 170 Al: 50 ZrO2: 165.5 Ag: 50 total film thickness: 1515 nm
Ti–Ag–ZrO2 Ti–Bragg
TiO2 SiO2
Ti–Ti–Al–ZrO2
Ti–Ti–Ag–ZrO2
ZrO2: Al: 50 Ti: ZrO2: 165.5 Ag: 50 Ti: 10
O2(g) [sccm]
Process pressure [mTorr]
Base pressure [mTorr]
Power [W]
40 40 10 40 10 40
7.0 7.0 0.0 7.0 0.0 6.5
2.0 2.0 5.0 2.0 5.0 2.0
1.0 0.6
348 346 200 338 250 398
100 40 10 100 40 10 100
3.5 7.0 0.0 0.0 7.0 0.0 0.0
5.0 2.0 5.0 5.0 2.0 5.0 5.0
Ar(g) [sccm]
0.9 1.2
0.6
0.9
300 346 200 325 336 250 325
journal of the mechanical behavior of biomedical materials 47 (2015) 1–11
Fig. 2 – Schematic representation of the geometric parameters of the two models to evaluate the fracture toughness Kc. (a) In the classical model, the crack length c of the radial cracks, emanating from the corners of the indent, is measured and the mean value is used for each indent. (b) The circumference s of the cracking around the indent was measured to calculate Afrac for the energy-based model. For these calculations, it was assumed that all cracks reach through the entire film thickness d.
depth, then decreased to a constant value. Coating systems with and without adhesion layer exhibited similar H resp. Er values. For ZrO2 based coatings, H reached around 9 GPa and Er around 120 GPa at a depth of 200 nm. At indentation depths below 100 nm, the Ti–ZrO2 sample showed higher H resp. Er values, followed by the systems with Al (Ti–Al–ZrO2 and Ti–Ti–Al–ZrO2) resp Ag layers (Ti–Ag–ZrO2 and Ti–Ti–Ag– ZrO2). The Bragg coated samples (Ti–Bragg, Ti–TiO2–Bragg) showed for both properties 50% lower values compared to the ZrO2 based systems. The substrate material showed almost matching H values as Bragg coated samples, but differed strongly from it for Er values. In Fig. 4, the results are presented after indentation to higher loads (10 to 50 mN). In general, only Bragg coated samples showed radial cracks emanating from the three corners of the indent. Circumferential cracking at the edges of the indent dominated for the remaining coating systems. Radial cracks were often not observed at the corners of the indent, although they were present at the edges. Furthermore, it was observed that the entire circumference of the indent cracked up for Ti–Ag–ZrO2 samples. Compared to the corresponding coating system with adhesion layer Ti–Ti–Ag– ZrO2, there were fewer cracks present on the surface and around the indent. Similarly, less cracking was also observed for Ti–Ti–Al–ZrO2 compared to Ti–Al–ZrO2 samples. At 40 and 50 mN, the coating bulged and partially spalled for Ti–Ag– ZrO2 samples.
4.2.
FIB cross sections
Cross sections were made at the positions indicated in Fig. 4 and the results are illustrated in Fig. 5. The Ti–ZrO2 sample showed that the ZrO2 coating was intact after indentation at 10 mN except for circumferential cracks in the coating, which were not going through the entire film. Extensive cohesive failure was observed for Ti–Al–ZrO2 and Ti–Ag–ZrO2 samples after indentation at 10 mN. Both metallic layers detached from the substrate material rather than from the ZrO2 coating. The delaminated area was also visible in the top view for
5
Ti–Ag–ZrO2 samples, where a circular pattern of dark and bright contrast surrounded the impression area (see inset in Fig. 5). Furthermore, the Ag layer deformed strongly and was pushed away from the apex towards the edge of the indent. This was not observed for the Al layer at the same load. Cross sections through Ti–Bragg showed that there might be delamination between the coating and the substrate material at a load of 30 mN. At this load, there was also shear cracking observed in the multilayer, which was not present at 10 mN. With increasing load, the length of the crack at the corners of the indent seemed to increase and propagated through the film and towards the substrate. Note that the cracks slightly deflected at each interface of the multilayer coating and did not reach to or into the substrate. The apparent crack feature in Ti was due to a curtaining artifact from the FIB milling process (Giannuzzi et al., 2005). Coating systems with an adhesion layer did not show any cohesive failure or delamination, though, circumferential cracks were still observed in the ZrO2 coating. Shear cracking in the multilayer was observed to a lesser extent in Ti–TiO2– Bragg samples after indentation at 30 mN.
4.3.
Calculation of the fracture toughness kc
The FIB cross sectioning elucidated the fracture scenario after indentation and, therefore, it was used to select the coating systems appropriate for the proposed Kc assessment models. In the following, the fracture toughness of Ti–ZrO2, Ti–Bragg and Ti–TiO2–Bragg coating systems were assessed using the models described before. The classical model can only be applied to brittle bulk materials. Since the Bragg based coating systems exhibited typical radial cracks at the corners, Eq. (2) was used to evaluate these samples. Furthermore, Kc of Ti–Bragg and Ti– TiO2–Bragg samples was calculated for loads Pm r30 mN as there was no substrate effect observed up to these loads. The crack length was measured according to Fig. 2(a) and the mean value was used for the calculation. The value for the empirical constant α was 0.04 for a cube corner indenter tip (Pharr, 1998). Instead of the Young’s modulus E, here, the reduced modulus Er of the samples was used in Eq. (2). Poisson’s ratio was not known for this multilayer system and according to the literature, it has a minor effect on the value of the Young modulus (Mesarovic and Fleck, 1999; Saha and Nix, 2002). The results for Ti–Bragg and Ti–TiO2–Bragg samples are presented in Table 2. The energy-based model was used to calculate Kc for Ti–ZrO2 samples after indentation at 10 mN and 20 mN. The fracture area Afrac of each indent was determined from the SEM image, while it was assumed that cracks were developed through the entire film thickness: Afrac ¼ s d
ð9Þ
With s as the circumference of the cracks surrounding the impression and d as the film thickness (see Fig. 2(b)). Additionally, the reduced modulus Er was taken from the results in Fig. 3, where Er E 109 GPa at a displacement of 45 nm. The results for Ti–ZrO2 are presented in Table 3.
6
journal of the mechanical behavior of biomedical materials 47 (2015) 1 –11
Fig. 3 – (a), (c) Hardness and (b), (d) reduced modulus of the polished Ti substrate and of all coating systems investigated here.
5.
Discussion
5.1.
Mechanical properties of the coating systems
Displacement-controlled experiments for ZrO2 based systems suggested that, with increasing indentation depth higher than 100 nm, hardness and reduced modulus decreased towards lower values and the effect of the substrate influences the measured values for further indentations at higher depth. Similar behavior of the curves (first increase then decrease of the values), was also observed for tungsten thin films on different substrates (Saha and Nix, 2002). At indentation depths smaller than 80–100 nm, the properties of the film should be displayed. Generally, the properties can be determined at indentation depths r10% of the film thickness as suggested in literature (Oliver and Pharr, 1992; Saha and Nix, 2002). This approach is feasible for films with a thickness in the mm-range, but not for very thin films as it was the case for the ZrO2-based coatings. In order to characterize the properties of the ZrO2 layer only in this study, an indentation depth of 16–17 nm would correspond to the 10% of the film thickness. Such depths impede the determination of the tip area function in this range since the shape of the indenter tip changes from a sharp three-sided pyramid to a spherically shaped tip (Oliver and Pharr, 1992). Additionally, the detection range of the machine also limited measurements for very small depths as the standard deviation increased strongly. Moreover, the surface roughness of the sample is likely to be another source of scattering for small indentation
depths. For the characterization of the ZrO2 layer here it was assumed that the H and Er values of the lowest possible indentation depth (45 nm) would represent the mechanical properties of the coating itself. The influence of the intermediate layers, Al resp. Ag, on the Er- and particularly H-curves can be also observed in Fig. 6. Due to the presence of the metallic layers, the mechanical properties were slightly increased for small indentation depths. Furthermore, the Ti–Ag–ZrO2 system showed higher H and Er values than the Ti–Al–ZrO2 system in this range. The Ag layer has a Young’s modulus of EAg ¼85–112 GPa and a hardness of HAg ¼ 1.5–1.6 GPa (Cao et al., 2008), whereas the Al layer has a Young’s modulus of EAl ¼75 GPa and a hardness of HAl ¼ 0.95 GPa (Saha and Nix, 2002), which might explain the observed behavior. For Ti–Bragg samples, H noticeably increased by almost 50% from a depth of 45 nm to 120 nm. A constant hardness of 6 GPa was reached for hZ120 nm. In the lower depth range, only the top layers contributed to the hardness. With increasing indentation depth, the influence of each layer on the hardness dropped since more layers were addressed during indentation and the properties of the entire multilayer can be observed for hZ120 nm. A similar behavior was also observed for the substrate material. The hardness increased as well by almost 50% in the same depth range. As mentioned previously, the shape of the tip differs for low indentation depths, which might influenced the measurement. Furthermore, this decrease may be also caused by a deformation layer on Ti, which developed, additionally to its natural TiO2 layer, during the grinding and polishing with a silica suspension (Petzow, 1999).
journal of the mechanical behavior of biomedical materials 47 (2015) 1–11
7
Fig. 4 – SEM images (all with BSE contrast) of indentations on the investigated samples at different loads (10–50 mN). With increasing load the residual indentation area increased for each sample. The maximum indentation depth hm is displayed above each image. The dotted lines indicate the position of the FIB cross sections.
5.2.
Adhesion between substrate and coating
No delamination or spallation occurred between the Ti substrate and the ZrO2 layer. Delamination was observed for the systems Ti–Al–ZrO2 and Ti–Ag–ZrO2, however. While good adhesion could be achieved in some systems if the deposition temperature is increased in order to facilitate diffusional intermixing between coating and substrate (Bull et al., 1991), in this study, the deposition was carried out at room temperature, to prevent a decrease in strength of the substrate material resp. the implant material for the future
application. Additionally, it is assumed that the substrate material still exhibits its natural oxide layer, which adversely affects the adhesion between titanium dioxide and inert metals in the systems Ti–Al–ZrO2 and Ti–Ag–ZrO2. However, other studies and our own showed that with the use of a Ti interlayer an improved adhesion can be achieved (Bull et al., 1991; Cheng et al., 1989; Rickerby and Burnett, 1988), as it has been observed in Fig. 5. The improved adhesion due to this additional layer can be explained by the chemical gettering effect between the Ti layer and the natural oxide layer of the substrate (Bull et al., 1991). The Ti–TiO2–Bragg samples
8
journal of the mechanical behavior of biomedical materials 47 (2015) 1 –11
Fig. 5 – FIB cross sections from all investigated samples at 10 mN and additionally at 30 mN for Ti–Bragg and Ti–TiO2–Bragg samples. The arrows indicate delaminated regions at the coating–substrate interface. Furthermore, shearing in load direction between the layers was observed and is indicated with two half arrows. The inset for Ti–Ag–ZrO2 (tested at 10 mN) shows a SE image, where a clear pattern of dark and bright contrast is visible around the impression. The positions of the cross sections are marked in Fig. 4. showed not only improvements in adhesion but also less shear effects in the coating itself. Presumably, the stresses at the interface in Ti–Bragg samples were increased, possibly due to a mismatch in elastic properties between the last layer of the coating SiO2 and the substrate Ti (ETiO2 ¼ 248–282 GPa, ESiO2 ¼ 60–70 GPa (Pelleg, 2012)).
5.3. Assessment of the fracture toughness for the investigated coatings After the analysis of the FIB cross sections, specific coating systems were selected for the assessment of Kc using the two models proposed earlier. The multilayer coating exhibited a high enough film thickness to be evaluated in a classical way (Eq. (2)). Furthermore, cross sections revealed the limit in the applied load, which did not generate a substrate effect. The Kc values for both Bragg coated samples (with and without adhesion layer) did not strongly differ from each other. It is therefore assumed that the classical model described here was appropriately applied for the Bragg coated samples. However, dental ceramics used for dental prosthesis offer a higher fracture toughness (KIc ¼6–10 MPa m1/2 (Christel et al., 1989)) than Ti–Bragg and Ti–TiO2–Bragg samples. The ZrO2 coated samples showed higher Kc values, which were also in the same range as for non-stabilized ZrO2 (2.6 MPa m1/2 (BravoLeon et al., 2002)). It is not apparent, though, that Kc increased by almost 40% when tested at 20 mN indentation for Ti–ZrO2 samples. There might be a load limit where the energy-based
model cannot be applied to such systems anymore due to overlapping fracture effects (e.g. cracking of the substrate) that are not considered in this model. Lower loads would be more acceptable for evaluation, but then the extent of cracking in the coating decreases as well. As the fracture area Afrac could not be determined for the coating systems with Al resp. Ag layer, the evaluation of Kc with the energy-based model for was not conducted. A solution to this issue would be to use FIB tomography for the determination of Afrac. With this method, the entire area of fracture, resulting from cracking and/or delamination, could be added up to one value. This procedure, however, is very time-consuming and additional image processing would be needed to determine this value accurately. Furthermore, the energy-based model described in this work can only be used as a semi-quantitative method. The assumption, which has been made for Ti–ZrO2 sample in Fig. 2(b), was not confirmed in the FIB cross sections. The cracks were only partially developed through the film thickness, and not entirely. Here, the approach using the FIB tomography method would be reasonable as well. However, an overestimation of Afrac, using a crack through the entire film thickness, gave at least the lower threshold of the fracture toughness, and therefore, even higher toughness values can be expected from these samples. Another possibility to compare the results of this study would be to use another substrate material, e.g. float glass substrates (Chen and Bull, 2007), which is harder and stiffer than Ti and would also reduce the plastic deformation arising from the substrate. Improvements to increase the fracture toughness of the ZrO2-based coatings could be achieved by
journal of the mechanical behavior of biomedical materials 47 (2015) 1–11
9
Table 2 – Measured parameters (c, H and Er) and calculated fracture toughness Kc for Ti–Bragg and Ti–TiO2–Bragg samples at loads Pm of 10, 20 and 30 mN based on the classical model. Coating system
Load Pm [mN]
Mean crack length c [lm]
Hardness H [GPa]
Reduced modulus Er [GPa]
Fracture toughness Kc [MPa (m)1/2]
Ti–Bragg
10 20 30 10
1.4670.04 2.2770.07 3.1570.12 1.3670.05
8.270.2 8.470.2 7.870.3 8.770.2
83.371.2 92.571.9 91.672.7 84.771.1
0.7370.03 0.7870.03 0.7470.04 0.7970.04
20 30
2.1970.18 3.0970.09
8.970.2 8.470.2
94.371.9 96.672.5
0.8270.08 0.7470.03
Ti–TiO2– Bragg
Table 3 – Fracture energy Ufrac, critical strain energy release rate Gc and fracture toughness Kc for Ti–ZrO2 samples at loads Pm of 10 and 20 mN based on the energy-based model. Coating system
Load Pm [mN]
Reduced modulus Er (at 45 nm) [GPa]
Fracture energy Ufrac [pJ]
Critical strain energy release rate Gc [N m/m2]
Fracture toughness Kc [MPa m1/2]
Ti–ZrO2
10 20
10973.9
20079.6 635782.7
69715 131727.5
2.7270.30 3.7770.40
doping yttrium oxide (Y2O3) to ZrO2. A way to confirm the validity of the semi-quantitative evaluation method would be to conduct the same experiments on samples coated with a Y2O3 doped ZrO2 layers of the same thickness and sputter parameters. If similar Kc values as for dental materials (up to 10 MPa m1/2, (Christel et al., 1989)) would be achieved, then this would approve the validity of the proposed model.
6.
Conclusions
In this study, coating systems, which improved the esthetical appearance of Ti (Pecnik et al., 2014a, 2014b), were characterized using nanoindentation. The characterization steps were conducted in a qualitative and semi-quantitative way. It can be concluded that:
With regards to the resistance against contact damage, thick
Fig. 6 – (a) Hardness and (b) reduced modulus of ZrO2-based coating systems without adhesion layer versus the normalized indentation depth (to illustrate the effect of the intermediate layers Al and Ag).
coatings shield the substrate from plastic deformation and fail in a brittle manner. Thin coatings fail as a consequence of the plastic deformation of the substrate. The addition of optically improving metallic layers (Al, Ag) requires the introduction of a Ti adhesion layer. The fracture toughness has been evaluated using two different approaches.
For coatings with thicknesses in the μm-range, their fracture toughness can be calculated applying the classical model. Another existing model, considering the substrate effect, was further modified for thin films and thin multilayer coatings on Ti substrates. However, the model should be treated in a semi-quantitative way as the assumed fracture profile did not correspond to the observed one.
10
journal of the mechanical behavior of biomedical materials 47 (2015) 1 –11
Acknowledgments The authors would like to thank the Scientific Center for Optical and Electron Microscopy (ScopeM) of ETH Zürich for the use of the Helios FIB SEM and the Laboratory of Metal Physics and Technology (LMPT) for the use of the Hitachi SEM. The present study was a part of a multidisciplinary research project granted by the Competence Centre for Material Science and Technology (CCMX, project no. 43) within the framework of the project “Colored Ceramic Surfaces for Metallic Dental Implants and Prosthetic Appliances”. The research partners were the Swiss Federal Institute of Technology Zurich (ETHZ), the University of Zurich (UZH), the Swiss Federal Laboratories for Material Science and Technology (Empa) and the Institut Straumann AG.
references
Albrektsson, T., Bra˚nemark, P.I., Hansson, H.A., Lindstrom, J., 1981. Osseointegrated titanium implants—requirements for ensuring a long-lasting, direct bone-to-implant anchorage in man. Acta Orthop. Scand. 52, 155–170. Andersson, B., 1995. Implants for single-tooth replacement. A clinical and experimental study on the Bra˚nemark CeraOne System. Swedish Dent. J. Suppl. 108, 1–41. Bra˚nemark, P., Hansson, B., Adell, R., Breine, U., Lindstro¨m, J., Halle´n, O., Ohman, A., 1977. Osseointegrated implants in the treatment of the edentulous jaw. Experience from a 10-year period. Scand. J. Plast. Reconstruct. Surg. Suppl. 16, 1. Bravo-Leon, A., Morikawa, Y., Kawahara, M., Mayo, M.J., 2002. Fracture toughness of nanocrystalline tetragonal zirconia with low yttria content. Acta Mater. 50, 4555–4562. Bull, S.J., Chalker, P.R., Ayres, C.F., Rickerby, D.S., 1991. The influence of titanium interlayers on the adhesion of titanium nitride coatings obtained by plasma-assisted chemical vapordeposition. Mater. Sci. Eng. A 139, 71–78. Bull, S.J., 2011. Analysis methods and size effects in the indentation fracture toughness assessment of very thin oxide coatings on glass. Comptes Rendus Mecanique. 339, 518–531. Buser, D., Janner, S.F.M., Wittneben, J.G., Bragger, U., Ramseier, C.A., Salvi, G.E., 2012. 10-year survival and success rates of 511 titanium implants with a sandblasted and acid-etched surface: a retrospective study in 303 partially edentulous patients. Clin. Implant Dent. Relat. Res. 14, 839–851. Cao, Y.F., Allameh, S., Nankivil, D., Sathiaraj, T.S., Otiti, T., Soboyejo, W., 2008. Nanoindentation measurements of the mechanical properties of polycrystalline Au and Ag thin films on silicon substrates: effects of grain size and film thickness (vol 427, pg 232, 2006). Mater. Sci. Eng. A 494, 466 –466. Chen, J., Bull, S.J., 2006. Assessment of the toughness of thin coatings using nanoindentation under displacement control. Thin Solid Films. 494, 1–7. Chen, J., Bull, S.J., 2007. Indentation fracture and toughness assessment for thin optical coatings on glass. J. Phys. D-Appl. Phys. 40, 5401–5417. Chen, J.J., Bull, S.J., 2011. Approaches to investigate delamination and interfacial toughness in coated systems: an overview. J. Phys. D-Appl. Phys., 44. Chen, J.J., 2012. Indentation-based methods to assess fracture toughness for thin coatings. J. Phys. D—Appl. Phys., 45.
Cheng, C.C., Erdemir, A., Fenske, G.R., 1989. Correlation of interface structure with adhesive strength of ion-plated TiN hard coatings. Surf. Coat. Technol. 39–40 (Part 1), 365–376. Cheng, Y.T., Li, Z.Y., Cheng, C.M., 2002. Scaling relationships for indentation measurements. Philos. Mag. A 82, 1821–1829. Christel, P., Meunier, A., Heller, M., Torre, J.P., Peille, C.N., 1989. Mechanical properties and short-term in-vivo evaluation of yttrium-oxide-partially-stabilized zirconia. J. Biomed. Mater. Res. 23, 45–61. Cook, R.F., Pharr, G.M., 1990. Direct observation and analysis of indentation cracking in glasses and ceramics. J. Am. Ceram. Soc. 73, 787–817. Dierens, M., Vandeweghe, S., Kisch, J., Nilner, K., De Bruyn, H., 2012. Long-term follow-up of turned single implants placed in periodontally healthy patients after 16–22 years: radiographic and peri-implant outcome. Clin. Oral Implants Res. 23, 197–204. Doerner, M.F., Nix, W.D., 1986. A method for interpreting the data from depth-sensing indentation instruments. J. Mater. Res. 1, 601–609. Gerberich, W.W., Cordill, M.J., 2006. Physics of adhesion. Rep. Prog. Phys. 69, 2157–2203. Giannuzzi, L.A., Kempshall, B., Schwarz, S., Lomness, J., Prenitzer, B., Stevie, F., 2005. FIB Lift-out Specimen Preparation Techniques (Introduction to Focused Ion Beams). Springer201–228. Harding, D.S., Oliver, W.C., Pharr, G.M., 1995. Cracking during nanoindentation and its use in the measurement of fracture toughness. Thin Films: Stresses Mech. Prop. V 356, 663–668. Joss-Vassalli, I., Grebenstein, C., Topouzelis, N., Sculean, A., Katsaros, C., 2010. Orthodontic therapy and gingival recession: a systematic review. Orthodont. Craniofacial Res. 13, 127–141. Jung, R.E., Sailer, I., Ha¨mmerle, C.H., Attin, T., Schmidlin, P., 2007. In vitro color changes of soft tissues caused by restorative materials. Int. J. Periodont. Restor. Dent. 27, 251–257. Lawn, B.R., Evans, A.G., Marshall, D.B., 1980. Elastic/plastic indentation damage in ceramics: the median/radial crack system. J. Am. Ceram. Soc. 63, 574–581. Lawn, B.R., Howes, V.R., 1981. Elastic recovery at hardness indentations. J. Mater. Sci. 16, 2745–2752. Lu, M.Y., Xie, H.T., Huang, H., 2013. Characterisation of interfacial adhesion of thin film/substrate systems using indentationinduced delamination: a focused review. Prog. Surf. Treat. Ii. 533, 201–222. Mesarovic, S.D., Fleck, N.A., 1999. Spherical indentation of elastic– plastic solids. Proc. R. Soc. London Ser. A 455, 2707–2728. Oliver, W.C., Pharr, G.M., 1992. An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments. J. Mater. Res.. 7, 1564–1583. Pecnik, C.M., Roos, M., Muff, D., Spolenak, R., Sailer, I., 2014a. In vitro color evaluation of esthetic coatings for metallic dental implants and implant prosthetic appliances. Clin. Oral Impl. Res.http: //dxdoi.org/10.1111/clr.12443 First published online: 15.07.2014. Pecnik, C.M., Muff, D., Spolenak, R., Sailer, I., 2014b. Color evaluation of a dielectric mirror coating using porcine tissue and prosthetic gingival material: a comparison of two models. Clin. Oral Impl. Res.. Pelleg, J., 2012. Mechanical Properties of Materials. Springer. Petzow, G., 1999. Metallographic Etching: Techniques for Metallography, Ceramography, Plastography, second ed. ASM International. Pharr, G.M., 1998. Measurement of mechanical properties by ultra-low load indentation. Mater. Sci. Eng. A 253, 151–159. Rickerby, D.S., Burnett, P.J., 1988. Correlation of process and system parameters with structure and properties of physically vapourdeposited hard coatings. Thin Solid Films. 157, 195–222. Ro¨sler, J., Harders, H., Ba¨ker, M., 2007. Mechanical Behaviour of Engineering Materials: Metals, Ceramics, Polymers, and Composites. Springer, Berlin; New York.
journal of the mechanical behavior of biomedical materials 47 (2015) 1–11
Saha, R., Nix, W.D., 2002. Effects of the substrate on the determination of thin film mechanical properties by nanoindentation. Acta Mater. 50, 23–38. Schiffmann, K.I., 2011. Determination of fracture toughness of bulk materials and thin films by nanoindentation: comparison of different models. Philos. Mag. A 91, 1163–1178.
11
van Brakel, R., Noordmans, H.J., Frenken, J., de Roode, R., de Wit, G.C., Cune, M.S., 2011. The effect of zirconia and titanium implant abutments on light reflection of the supporting soft tissues. Clin. Oral Impl. Res. 22, 1172–1178.
Available online at www.sciencedirect.com
www.elsevier.com/locate/jmbbm
Research Paper
Fracture toughness of esthetic dental coating systems by nanoindentation and FIB sectional analysis Christina Martina Pecnik, Diana Courty, Daniel Muff, Ralph Spolenakn Laboratory for Nanometallurgy, Department of Materials, ETH Zurich, Zurich, Switzerland
art i cle i nfo
ab st rac t
Article history:
Improving the esthetics of Ti-based dental implants is the last challenge remaining in the
Received 10 November 2014
optimization process. The optical issues were recently solved by the application of highly
Received in revised form
and selectively reflective coatings on Ti implants. This work focuses on the mechanical
28 February 2015
durability of these esthetic ceramic based coating systems (with and without adhesion
Accepted 10 March 2015
layers).
Available online 19 March 2015
The coating systems (Ti–ZrO2, Ti–Al–ZrO2, Ti–Ti–Al–ZrO2, Ti–Ag–ZrO2, Ti–Ti–Ag–ZrO2,
Keywords:
Ti–Bragg and Ti–TiO2–Bragg) were subjected to nanoindentation experiments and exam-
Fracture toughness
ined using scanning electron microscopy and focused ion beam cross sectional analysis.
Dental implants
Three coating systems contained adhesion layers (10 nm of Ti or 60 nm of TiO2 layers). The
Thin films
fracture toughness of selected samples was assessed applying two different models from
Nanoindentation
literature, a classical for bulk materials and an energy-based model, which was further
FIB
developed and adjusted.
Titanium Ceramics
The ZrO2 based coating systems (total film thicknesso200 nm) followed a circumferential cracking behavior in contrast to Bragg coated samples (total film thickness around 1.5 μm), which showed radial cracking emanating from the indent corners. For Ti–ZrO2 samples, a fracture toughness between 2.70 and 3.70 MPa m1/2 was calculated using an energy-based model. The classical model was applied to Bragg coated samples and their fracture toughness ranged between 0.70 and 0.80 MPa m1/2. Furthermore, coating systems containing an additional layer (Ti–Ti–Al–ZrO2, Ti–Ti–Ag–ZrO2 and Ti–TiO2-Bragg) showed an improved adhesion between the substrate and the coating. The addition of a Ti or TiO2 layer improved the adhesion between substrate and coating. The validity of the models for the assessment of the fracture toughness depended on the layer structure and fracture profile of the samples investigated here (classical model for thick coatings and energy-based model for thin coatings). & 2015 Elsevier Ltd. All rights reserved.
n Correspondence to: Laboratory for Nanometallurgy, Department of Materials, ETH Zurich, Vladimir-Prelog-Weg 5, 8093 Zurich, Switzerland. Tel.: þ41 44 632 25 90; fax: þ41 44 632 11 01. E-mail address: [email protected] (C.M. Pecnik).
http://dx.doi.org/10.1016/j.jmbbm.2015.03.006 1751-6161/& 2015 Elsevier Ltd. All rights reserved.
2
1.
journal of the mechanical behavior of biomedical materials 47 (2015) 1 –11
Introduction
Dental implants have revolutionized the reconstructive dentistry since the late 1970’s and offer a reliable treatment to reestablish a healthy and functional dentition after tooth-loss (Brånemark et al., 1977). Titanium (Ti) is still considered as the ‘gold-standard’ for dental reconstructions, since it exhibits a high fatigue and corrosion resistance, biocompatibility and excellent osseointegration as shown in long-term studies (Albrektsson et al., 1981; Andersson, 1995; Buser et al., 2012; Dierens et al., 2012). An important aspect, though, is the esthetical appearance of the peri-implant soft tissue: metallic reconstructions can negatively influence the esthetics caused by recession and/or dark discolorations of the gingival tissue and, therefore, impair the patient's approval of the treatment’s results (Joss-Vassalli et al., 2010; Jung et al., 2007; van Brakel et al., 2011). This disadvantage of the Ti implant can be remedied by ceramic based coatings (Pecnik et al., 2014a, 2014b), which showed significantly improved optical properties compared to Ti and were suggested as esthetic dental coatings for future application. In addition to the optical analysis, a mechanical characterization of these thin film systems is needed since their performance should be reliable during service. One simple way to assess the mechanical properties of thin films is by nanoindentation (Doerner and Nix, 1986; Oliver and Pharr, 1992). For thin films, the hardness still can be determined at small indentation depths (o10% of the film thickness) (Saha and Nix, 2002), whereas the assessment of the fracture toughness becomes a challenge. Various models on fracture mechanisms in thin films have been described in literature; however, the use of the appropriate method depends on the thickness, cracking scenario and shape of the load–displacement curve of the tested film (Bull, 2011; Chen and Bull, 2007, 2011; Schiffmann, 2011). Despite the fact that sharp indentation is infrequently occurring in the periimplant region respectively in the soft tissue part of the implant, nanoindentation is a convenient technique to measure not only the material’s properties, but it gives also information about the adhesion and/or cohesion of coatings (Chen and Bull, 2011; Gerberich and Cordill, 2006; Lu et al., 2013). This method allows to characterize the fracture toughness of thin films under clearly defined loading conditions, contrary to the complex actual loading scenarios present in the oral environment The fracture toughness of a material is not only a property that describes the resistance of further cracking, but it is also helpful to value its reliability under mechanical stress. Nanoindentation with a cube corner tip was conducted at different loads to create cracking in the films. Furthermore, focused ion beam (FIB) cross sections of the samples were analyzed in order to reveal the extent of fracture and deformation in the film, substrate and in between them after indentation. Some of the investigated coating systems exhibited very thin film thicknesses (below 200 nm), which is a challenging task for the implementation of nanoindentation experiments and their analysis. Accordingly, different models such as the classical model used for bulk materials (Lawn et al., 1980) and energy-based models by Bull and Chen (2011), Chen and Bull (2007) were applied to calculate the fracture toughness of the coating systems. Furthermore, selected
coating systems contained adhesion layers, which ultimately provided information about the interplay between coating and substrate visible through FIB cross sections. Moreover, the purpose of these experiments was also to compare the results qualitatively between the coating systems and finally, discuss the validity of the quantitative results and calculations from the applied models.
2. Overview of models for the determination of fracture toughness by nanoindentation The determination of the fracture toughness of very thin films (o500 nm) is a difficult task for which, in the last 30 years, different models and techniques were developed. In general, the fracture toughness for mode I can be described by the following equation: pffiffiffiffiffiffiffiffi ð1Þ KIc ¼ EGc where KIc is also called the critical stress-intensity factor, E is the Young modulus and Gc is the critical strain energy release rate, which corresponds to the dissipated energy of a crack per unit area. Mode I describes the tensile loading acting perpendicularly on a crack (Rösler et al., 2007). After indentation in a brittle material, there are different kinds of crack initiation patterns and stress fields. Most applied crack geometries are halfpenny-shaped and Palmquist (or radial) cracks for brittle materials, which were classified by Cook and Pharr (1990). It is, however, difficult to distinguish between crack geometries from the top view of an indent. Mainly, the analysis of the half-penny crack system is used to determine the toughness of brittle materials (Bull, 2011). On the assumption of the half-penny crack system Lawns et al. developed an equation to determine the fracture toughness after indentation by measuring the length of cracks that originate at the corners of the indent (Lawn et al., 1980): 1=2 E P 3=2 ð2Þ Kc ¼ α H c where Kc is now addressed to the fracture toughness since the loading condition in indentation experiments does not correspond to a pure mode I loading. H is the hardness, P is the applied load, and c is the length of the crack emanating from the corners of the indent. The empirical constant α depends on the indenter geometry and crack system, but there is a range of different values in literature for each tip geometry and crack pattern, which will be discussed later on. This equation, however, can only be applied for bulk brittle materials or sufficiently thick brittle films, as suggested by Bull (2011), Chen (2012). In contrast to the conventional, classical indentation methods, which are more suitable for bulk materials, nanoindentation is a well-established method to determine the Young modulus and hardness of thin films. For fracture toughness experiments, one has to facilitate the creation of cracks at the corners of the indent. Suitably, the cube corner tip exhibits sharp sides that induce high enough stress intensities and, therefore, even lowers the threshold of load for cracking (Harding et al., 1995; Pharr, 1998). As the coating thickness decreases, further problems could arise for the assessment of the fracture toughness: radial cracking may no longer be observed and with increasing load,
journal of the mechanical behavior of biomedical materials 47 (2015) 1–11
the effect from the substrate becomes more significant. Alternative methods have been developed by Bull (2011), Chen and Bull (2006, 2007, 2011), Chen (2012), where energy-based models were applied for the determination of the fracture toughness for very thin films. In their alternative approach the total work of indentation (Wtot) corresponded to the area below the load– displacement (P–h) curve, which can be further split in different parts of the total work such as elastic, plastic and fracture part. Depending on the presence of features in the P–h curve, two different methods are described to determine the fracture toughness. Here, the model without features (e.g. no slope change or discontinuities) in the P–h curve will be described in detail.
2.1.
Energy-based model for P–h curves without features
The total work of indentation can be defined as (Bull, 2011; Chen and Bull, 2007) Wtot ¼ Welast þ Wplast þ Ufrac þ Wother
ð3Þ
where Welast is the work of elastic deformation, Wplast is the work of plastic deformation, Ufrac is the dissipated energy during fracture, and Wother is the work due to heat dissipation and/or creep during indentation. Wother will be neglected in the following since it is assumed to be a small term contributing to the total work. This then reduces Eq. (3) to the following equation: Wtot ¼ Welast þ Wplast þ Ufrac
ð4Þ
Another energy term can be defined, which consists of the plastic and fracture energy: Uirr ¼ Wplast þ Ufrac
ð5Þ
where Uirr is the irreversible energy, whereas the reversible
3
plastic behavior is not considered in this case. After an indentation experiment, Wtot, Welast, Uirr can directly be obtained from the resulting load displacement curve. However, the plastic work Wplast has to be determined using a different approach (Cheng et al., 2002; Lawn and Howes, 1981): 2 3 h
2
h
f f 61 3 hm þ 2 hm Wplast ¼ 1 6 2 4 Wtot h 1 hmf
3
7 7 5
ð6Þ
The obtained fracture energy Ufrac can then be further used to define the critical strain energy release rate Gc: Gc ¼
Ufrac Afrac
ð7Þ
where Afrac is the area produced during cracking. Both interfacial and coating cracking can be considered in the total fracture area Afrac. Finally, the fracture toughness can be calculated using Eq. (1) with the substitution of KIc with Kc and E with Er: pffiffiffiffiffiffiffiffiffiffi ð8Þ Kc ¼ Er Gc
3.
Materials and methods
3.1.
Coating systems
The different coating systems in this study are illustrated in Fig. 1 and is based on the design strategy described in Pecnik et al. (2014a, 2014b). Additionally, some coating systems were deposited with an adhesion layer (Ti or TiO2). The deposition parameters are displayed in Table 1.
Fig. 1 – Layer structure of the coating systems. Seven coating systems were characterized in this study: all were deposited at a process chamber pressure of 2 mTorr. Additionally, the three samples contained adhesion layers (10 nm of Ti for ZrO2 based coating systems with Al resp. Ag and 60 nm of TiO2 for the Bragg coating).
4
journal of the mechanical behavior of biomedical materials 47 (2015) 1 –11
3.2.
Substrate preparation and thin film deposition
Round Ti platelets (Grade 4, ThyssenKrupp) were used as substrate material, and exhibited a thickness of 0.5 mm and a diameter of 15 mm. An automatic grinding and polishing machine (Rotopol-2, Struers) was used for substrate preparation. Initially, the platelets were ground using silicon carbide papers up to 4000 grid. Prior to the polishing step, the platelets were ultrasonically cleaned in ethanol for 5 min. Subsequently, the platelets were polished with a modified 0.02 μm colloidal silica suspension (MasterMet 2, Buehler), containing 10 vol% hydrogen peroxide (Z35%, Sigma-Aldrich). The polished substrates were ultrasonically cleaned with ethanol and acetone (1:1 mixture) for 10 min. The average area roughness of the polished Ti platelets was Sa ¼ 0.005 μm, obtained by atomic force microscopy (Cypher, Asylum Research, USA) (Pecnik et al., 2014a, 2014b). The coating systems (Table 1) were then deposited by reactive magnetron sputtering at room temperature (PVD Products, Inc.). The base pressure in the deposition chamber was 2 mTorr. The thickness of the reactively deposited zirconia (ZrO2) layer was selected for a pink interference color (165.5 nm). Both metallic intermediate layers, aluminum (Al, 99.9%, Sindlhauser Materials GmbH) and silver (Ag, 99.99%, Kurt J. Lesker), exhibited a thickness of 50 nm. Furthermore, an adhesion layer was deposited on three further Ti samples (10 nm of Ti as adhesion layer for ZrO2 based coatings with Al resp. Ag layers, and 10 nm of TiO2 as adhesion layer for the Bragg coating, see Fig. 1).
3.3.
Nanoindentation
The mechanical properties of the substrate and different coating systems were characterized using an Ultra Nanoindentation Tester (UNHT, CSM Instruments by Anton Paar, Switzerland) with a cube corner tip. The Young’s modulus Ei of the indenter tip was 1140 GPa and its Poisson’s ratio νi was 0.07. Displacement-controlled experiments were conducted for 12 different indentation depths (45 to 200 nm in different step sizes). For each indentation depth, the indentation was repeated 20 times with a distance of 10 μm between the indents. Loading and unloading rates were adjusted to the corresponding indentation depth (e.g. depth¼45 nm, loading
resp. unloading rate¼ 450 nm/min) without dwell time keeping the loading rate constant. The results were analyzed using the method by Oliver and Pharr (1992) to determine the hardness H and reduced modulus Er. Load-controlled experiments were conducted for five different loads (10 to 50 mN with 10 mN step size). For each load, the indentation was repeated 20 times with a distance of 30 μm between the indents. Loading and unloading rates were adjusted to the corresponding load (e.g. load¼ 10 mN, loading resp. unloading rate¼ 12 ∙ load [mN/min]¼120 mN/min) without dwell time keeping the loading rate constant.
3.4.
Characterization and evaluation
Indents from load-controlled experiments were analyzed by means of secondary electron imaging (SE) and backscattered electron imaging (BSE) using a Schottky-type field-emission scanning electron microscope (SU 70, Hitachi, Japan). Focused ion beam (FIB, Helios Nanolab 600i, FEI) technique was used to investigate the film cross sections after indentation. A platinum (Pt) layer was deposited on the indent in order to protect the sample surface. After cross-sectioning, the samples were analyzed using the featured Schottky-type fieldemission gun (Helios Nanolab 600i, FEI). The fracture toughness of each sample was evaluated using the previously described models for different loads. Fig. 2 shows the evaluation procedures for both models. Further details to the evaluation procedure will be described in the results section.
4.
Results
4.1.
Indentation results
The mechanical properties of the coating systems and t10he substrate material are shown in Fig. 3. The indentation depths were chosen in a rather small displacement range in order to characterize the properties of each coating itself, neglecting the impact of the substrate. Hardness and reduced modulus values first increased with increasing indentation
Table 1 – Sputter deposition parameters of the coating systems. For Bragg-based coatings, the design wavelength was λ1 ¼480 nm for the first stack (S1) and λ2 ¼590 nm for the second stack (S2). Sample description
Layer thickness [nm]
Ti–ZrO2 Ti–Al–ZrO2
ZrO2: 165.5 ZrO2: 170 Al: 50 ZrO2: 165.5 Ag: 50 total film thickness: 1515 nm
Ti–Ag–ZrO2 Ti–Bragg
TiO2 SiO2
Ti–Ti–Al–ZrO2
Ti–Ti–Ag–ZrO2
ZrO2: Al: 50 Ti: ZrO2: 165.5 Ag: 50 Ti: 10
O2(g) [sccm]
Process pressure [mTorr]
Base pressure [mTorr]
Power [W]
40 40 10 40 10 40
7.0 7.0 0.0 7.0 0.0 6.5
2.0 2.0 5.0 2.0 5.0 2.0
1.0 0.6
348 346 200 338 250 398
100 40 10 100 40 10 100
3.5 7.0 0.0 0.0 7.0 0.0 0.0
5.0 2.0 5.0 5.0 2.0 5.0 5.0
Ar(g) [sccm]
0.9 1.2
0.6
0.9
300 346 200 325 336 250 325
journal of the mechanical behavior of biomedical materials 47 (2015) 1–11
Fig. 2 – Schematic representation of the geometric parameters of the two models to evaluate the fracture toughness Kc. (a) In the classical model, the crack length c of the radial cracks, emanating from the corners of the indent, is measured and the mean value is used for each indent. (b) The circumference s of the cracking around the indent was measured to calculate Afrac for the energy-based model. For these calculations, it was assumed that all cracks reach through the entire film thickness d.
depth, then decreased to a constant value. Coating systems with and without adhesion layer exhibited similar H resp. Er values. For ZrO2 based coatings, H reached around 9 GPa and Er around 120 GPa at a depth of 200 nm. At indentation depths below 100 nm, the Ti–ZrO2 sample showed higher H resp. Er values, followed by the systems with Al (Ti–Al–ZrO2 and Ti–Ti–Al–ZrO2) resp Ag layers (Ti–Ag–ZrO2 and Ti–Ti–Ag– ZrO2). The Bragg coated samples (Ti–Bragg, Ti–TiO2–Bragg) showed for both properties 50% lower values compared to the ZrO2 based systems. The substrate material showed almost matching H values as Bragg coated samples, but differed strongly from it for Er values. In Fig. 4, the results are presented after indentation to higher loads (10 to 50 mN). In general, only Bragg coated samples showed radial cracks emanating from the three corners of the indent. Circumferential cracking at the edges of the indent dominated for the remaining coating systems. Radial cracks were often not observed at the corners of the indent, although they were present at the edges. Furthermore, it was observed that the entire circumference of the indent cracked up for Ti–Ag–ZrO2 samples. Compared to the corresponding coating system with adhesion layer Ti–Ti–Ag– ZrO2, there were fewer cracks present on the surface and around the indent. Similarly, less cracking was also observed for Ti–Ti–Al–ZrO2 compared to Ti–Al–ZrO2 samples. At 40 and 50 mN, the coating bulged and partially spalled for Ti–Ag– ZrO2 samples.
4.2.
FIB cross sections
Cross sections were made at the positions indicated in Fig. 4 and the results are illustrated in Fig. 5. The Ti–ZrO2 sample showed that the ZrO2 coating was intact after indentation at 10 mN except for circumferential cracks in the coating, which were not going through the entire film. Extensive cohesive failure was observed for Ti–Al–ZrO2 and Ti–Ag–ZrO2 samples after indentation at 10 mN. Both metallic layers detached from the substrate material rather than from the ZrO2 coating. The delaminated area was also visible in the top view for
5
Ti–Ag–ZrO2 samples, where a circular pattern of dark and bright contrast surrounded the impression area (see inset in Fig. 5). Furthermore, the Ag layer deformed strongly and was pushed away from the apex towards the edge of the indent. This was not observed for the Al layer at the same load. Cross sections through Ti–Bragg showed that there might be delamination between the coating and the substrate material at a load of 30 mN. At this load, there was also shear cracking observed in the multilayer, which was not present at 10 mN. With increasing load, the length of the crack at the corners of the indent seemed to increase and propagated through the film and towards the substrate. Note that the cracks slightly deflected at each interface of the multilayer coating and did not reach to or into the substrate. The apparent crack feature in Ti was due to a curtaining artifact from the FIB milling process (Giannuzzi et al., 2005). Coating systems with an adhesion layer did not show any cohesive failure or delamination, though, circumferential cracks were still observed in the ZrO2 coating. Shear cracking in the multilayer was observed to a lesser extent in Ti–TiO2– Bragg samples after indentation at 30 mN.
4.3.
Calculation of the fracture toughness kc
The FIB cross sectioning elucidated the fracture scenario after indentation and, therefore, it was used to select the coating systems appropriate for the proposed Kc assessment models. In the following, the fracture toughness of Ti–ZrO2, Ti–Bragg and Ti–TiO2–Bragg coating systems were assessed using the models described before. The classical model can only be applied to brittle bulk materials. Since the Bragg based coating systems exhibited typical radial cracks at the corners, Eq. (2) was used to evaluate these samples. Furthermore, Kc of Ti–Bragg and Ti– TiO2–Bragg samples was calculated for loads Pm r30 mN as there was no substrate effect observed up to these loads. The crack length was measured according to Fig. 2(a) and the mean value was used for the calculation. The value for the empirical constant α was 0.04 for a cube corner indenter tip (Pharr, 1998). Instead of the Young’s modulus E, here, the reduced modulus Er of the samples was used in Eq. (2). Poisson’s ratio was not known for this multilayer system and according to the literature, it has a minor effect on the value of the Young modulus (Mesarovic and Fleck, 1999; Saha and Nix, 2002). The results for Ti–Bragg and Ti–TiO2–Bragg samples are presented in Table 2. The energy-based model was used to calculate Kc for Ti–ZrO2 samples after indentation at 10 mN and 20 mN. The fracture area Afrac of each indent was determined from the SEM image, while it was assumed that cracks were developed through the entire film thickness: Afrac ¼ s d
ð9Þ
With s as the circumference of the cracks surrounding the impression and d as the film thickness (see Fig. 2(b)). Additionally, the reduced modulus Er was taken from the results in Fig. 3, where Er E 109 GPa at a displacement of 45 nm. The results for Ti–ZrO2 are presented in Table 3.
6
journal of the mechanical behavior of biomedical materials 47 (2015) 1 –11
Fig. 3 – (a), (c) Hardness and (b), (d) reduced modulus of the polished Ti substrate and of all coating systems investigated here.
5.
Discussion
5.1.
Mechanical properties of the coating systems
Displacement-controlled experiments for ZrO2 based systems suggested that, with increasing indentation depth higher than 100 nm, hardness and reduced modulus decreased towards lower values and the effect of the substrate influences the measured values for further indentations at higher depth. Similar behavior of the curves (first increase then decrease of the values), was also observed for tungsten thin films on different substrates (Saha and Nix, 2002). At indentation depths smaller than 80–100 nm, the properties of the film should be displayed. Generally, the properties can be determined at indentation depths r10% of the film thickness as suggested in literature (Oliver and Pharr, 1992; Saha and Nix, 2002). This approach is feasible for films with a thickness in the mm-range, but not for very thin films as it was the case for the ZrO2-based coatings. In order to characterize the properties of the ZrO2 layer only in this study, an indentation depth of 16–17 nm would correspond to the 10% of the film thickness. Such depths impede the determination of the tip area function in this range since the shape of the indenter tip changes from a sharp three-sided pyramid to a spherically shaped tip (Oliver and Pharr, 1992). Additionally, the detection range of the machine also limited measurements for very small depths as the standard deviation increased strongly. Moreover, the surface roughness of the sample is likely to be another source of scattering for small indentation
depths. For the characterization of the ZrO2 layer here it was assumed that the H and Er values of the lowest possible indentation depth (45 nm) would represent the mechanical properties of the coating itself. The influence of the intermediate layers, Al resp. Ag, on the Er- and particularly H-curves can be also observed in Fig. 6. Due to the presence of the metallic layers, the mechanical properties were slightly increased for small indentation depths. Furthermore, the Ti–Ag–ZrO2 system showed higher H and Er values than the Ti–Al–ZrO2 system in this range. The Ag layer has a Young’s modulus of EAg ¼85–112 GPa and a hardness of HAg ¼ 1.5–1.6 GPa (Cao et al., 2008), whereas the Al layer has a Young’s modulus of EAl ¼75 GPa and a hardness of HAl ¼ 0.95 GPa (Saha and Nix, 2002), which might explain the observed behavior. For Ti–Bragg samples, H noticeably increased by almost 50% from a depth of 45 nm to 120 nm. A constant hardness of 6 GPa was reached for hZ120 nm. In the lower depth range, only the top layers contributed to the hardness. With increasing indentation depth, the influence of each layer on the hardness dropped since more layers were addressed during indentation and the properties of the entire multilayer can be observed for hZ120 nm. A similar behavior was also observed for the substrate material. The hardness increased as well by almost 50% in the same depth range. As mentioned previously, the shape of the tip differs for low indentation depths, which might influenced the measurement. Furthermore, this decrease may be also caused by a deformation layer on Ti, which developed, additionally to its natural TiO2 layer, during the grinding and polishing with a silica suspension (Petzow, 1999).
journal of the mechanical behavior of biomedical materials 47 (2015) 1–11
7
Fig. 4 – SEM images (all with BSE contrast) of indentations on the investigated samples at different loads (10–50 mN). With increasing load the residual indentation area increased for each sample. The maximum indentation depth hm is displayed above each image. The dotted lines indicate the position of the FIB cross sections.
5.2.
Adhesion between substrate and coating
No delamination or spallation occurred between the Ti substrate and the ZrO2 layer. Delamination was observed for the systems Ti–Al–ZrO2 and Ti–Ag–ZrO2, however. While good adhesion could be achieved in some systems if the deposition temperature is increased in order to facilitate diffusional intermixing between coating and substrate (Bull et al., 1991), in this study, the deposition was carried out at room temperature, to prevent a decrease in strength of the substrate material resp. the implant material for the future
application. Additionally, it is assumed that the substrate material still exhibits its natural oxide layer, which adversely affects the adhesion between titanium dioxide and inert metals in the systems Ti–Al–ZrO2 and Ti–Ag–ZrO2. However, other studies and our own showed that with the use of a Ti interlayer an improved adhesion can be achieved (Bull et al., 1991; Cheng et al., 1989; Rickerby and Burnett, 1988), as it has been observed in Fig. 5. The improved adhesion due to this additional layer can be explained by the chemical gettering effect between the Ti layer and the natural oxide layer of the substrate (Bull et al., 1991). The Ti–TiO2–Bragg samples
8
journal of the mechanical behavior of biomedical materials 47 (2015) 1 –11
Fig. 5 – FIB cross sections from all investigated samples at 10 mN and additionally at 30 mN for Ti–Bragg and Ti–TiO2–Bragg samples. The arrows indicate delaminated regions at the coating–substrate interface. Furthermore, shearing in load direction between the layers was observed and is indicated with two half arrows. The inset for Ti–Ag–ZrO2 (tested at 10 mN) shows a SE image, where a clear pattern of dark and bright contrast is visible around the impression. The positions of the cross sections are marked in Fig. 4. showed not only improvements in adhesion but also less shear effects in the coating itself. Presumably, the stresses at the interface in Ti–Bragg samples were increased, possibly due to a mismatch in elastic properties between the last layer of the coating SiO2 and the substrate Ti (ETiO2 ¼ 248–282 GPa, ESiO2 ¼ 60–70 GPa (Pelleg, 2012)).
5.3. Assessment of the fracture toughness for the investigated coatings After the analysis of the FIB cross sections, specific coating systems were selected for the assessment of Kc using the two models proposed earlier. The multilayer coating exhibited a high enough film thickness to be evaluated in a classical way (Eq. (2)). Furthermore, cross sections revealed the limit in the applied load, which did not generate a substrate effect. The Kc values for both Bragg coated samples (with and without adhesion layer) did not strongly differ from each other. It is therefore assumed that the classical model described here was appropriately applied for the Bragg coated samples. However, dental ceramics used for dental prosthesis offer a higher fracture toughness (KIc ¼6–10 MPa m1/2 (Christel et al., 1989)) than Ti–Bragg and Ti–TiO2–Bragg samples. The ZrO2 coated samples showed higher Kc values, which were also in the same range as for non-stabilized ZrO2 (2.6 MPa m1/2 (BravoLeon et al., 2002)). It is not apparent, though, that Kc increased by almost 40% when tested at 20 mN indentation for Ti–ZrO2 samples. There might be a load limit where the energy-based
model cannot be applied to such systems anymore due to overlapping fracture effects (e.g. cracking of the substrate) that are not considered in this model. Lower loads would be more acceptable for evaluation, but then the extent of cracking in the coating decreases as well. As the fracture area Afrac could not be determined for the coating systems with Al resp. Ag layer, the evaluation of Kc with the energy-based model for was not conducted. A solution to this issue would be to use FIB tomography for the determination of Afrac. With this method, the entire area of fracture, resulting from cracking and/or delamination, could be added up to one value. This procedure, however, is very time-consuming and additional image processing would be needed to determine this value accurately. Furthermore, the energy-based model described in this work can only be used as a semi-quantitative method. The assumption, which has been made for Ti–ZrO2 sample in Fig. 2(b), was not confirmed in the FIB cross sections. The cracks were only partially developed through the film thickness, and not entirely. Here, the approach using the FIB tomography method would be reasonable as well. However, an overestimation of Afrac, using a crack through the entire film thickness, gave at least the lower threshold of the fracture toughness, and therefore, even higher toughness values can be expected from these samples. Another possibility to compare the results of this study would be to use another substrate material, e.g. float glass substrates (Chen and Bull, 2007), which is harder and stiffer than Ti and would also reduce the plastic deformation arising from the substrate. Improvements to increase the fracture toughness of the ZrO2-based coatings could be achieved by
journal of the mechanical behavior of biomedical materials 47 (2015) 1–11
9
Table 2 – Measured parameters (c, H and Er) and calculated fracture toughness Kc for Ti–Bragg and Ti–TiO2–Bragg samples at loads Pm of 10, 20 and 30 mN based on the classical model. Coating system
Load Pm [mN]
Mean crack length c [lm]
Hardness H [GPa]
Reduced modulus Er [GPa]
Fracture toughness Kc [MPa (m)1/2]
Ti–Bragg
10 20 30 10
1.4670.04 2.2770.07 3.1570.12 1.3670.05
8.270.2 8.470.2 7.870.3 8.770.2
83.371.2 92.571.9 91.672.7 84.771.1
0.7370.03 0.7870.03 0.7470.04 0.7970.04
20 30
2.1970.18 3.0970.09
8.970.2 8.470.2
94.371.9 96.672.5
0.8270.08 0.7470.03
Ti–TiO2– Bragg
Table 3 – Fracture energy Ufrac, critical strain energy release rate Gc and fracture toughness Kc for Ti–ZrO2 samples at loads Pm of 10 and 20 mN based on the energy-based model. Coating system
Load Pm [mN]
Reduced modulus Er (at 45 nm) [GPa]
Fracture energy Ufrac [pJ]
Critical strain energy release rate Gc [N m/m2]
Fracture toughness Kc [MPa m1/2]
Ti–ZrO2
10 20
10973.9
20079.6 635782.7
69715 131727.5
2.7270.30 3.7770.40
doping yttrium oxide (Y2O3) to ZrO2. A way to confirm the validity of the semi-quantitative evaluation method would be to conduct the same experiments on samples coated with a Y2O3 doped ZrO2 layers of the same thickness and sputter parameters. If similar Kc values as for dental materials (up to 10 MPa m1/2, (Christel et al., 1989)) would be achieved, then this would approve the validity of the proposed model.
6.
Conclusions
In this study, coating systems, which improved the esthetical appearance of Ti (Pecnik et al., 2014a, 2014b), were characterized using nanoindentation. The characterization steps were conducted in a qualitative and semi-quantitative way. It can be concluded that:
With regards to the resistance against contact damage, thick
Fig. 6 – (a) Hardness and (b) reduced modulus of ZrO2-based coating systems without adhesion layer versus the normalized indentation depth (to illustrate the effect of the intermediate layers Al and Ag).
coatings shield the substrate from plastic deformation and fail in a brittle manner. Thin coatings fail as a consequence of the plastic deformation of the substrate. The addition of optically improving metallic layers (Al, Ag) requires the introduction of a Ti adhesion layer. The fracture toughness has been evaluated using two different approaches.
For coatings with thicknesses in the μm-range, their fracture toughness can be calculated applying the classical model. Another existing model, considering the substrate effect, was further modified for thin films and thin multilayer coatings on Ti substrates. However, the model should be treated in a semi-quantitative way as the assumed fracture profile did not correspond to the observed one.
10
journal of the mechanical behavior of biomedical materials 47 (2015) 1 –11
Acknowledgments The authors would like to thank the Scientific Center for Optical and Electron Microscopy (ScopeM) of ETH Zürich for the use of the Helios FIB SEM and the Laboratory of Metal Physics and Technology (LMPT) for the use of the Hitachi SEM. The present study was a part of a multidisciplinary research project granted by the Competence Centre for Material Science and Technology (CCMX, project no. 43) within the framework of the project “Colored Ceramic Surfaces for Metallic Dental Implants and Prosthetic Appliances”. The research partners were the Swiss Federal Institute of Technology Zurich (ETHZ), the University of Zurich (UZH), the Swiss Federal Laboratories for Material Science and Technology (Empa) and the Institut Straumann AG.
references
Albrektsson, T., Bra˚nemark, P.I., Hansson, H.A., Lindstrom, J., 1981. Osseointegrated titanium implants—requirements for ensuring a long-lasting, direct bone-to-implant anchorage in man. Acta Orthop. Scand. 52, 155–170. Andersson, B., 1995. Implants for single-tooth replacement. A clinical and experimental study on the Bra˚nemark CeraOne System. Swedish Dent. J. Suppl. 108, 1–41. Bra˚nemark, P., Hansson, B., Adell, R., Breine, U., Lindstro¨m, J., Halle´n, O., Ohman, A., 1977. Osseointegrated implants in the treatment of the edentulous jaw. Experience from a 10-year period. Scand. J. Plast. Reconstruct. Surg. Suppl. 16, 1. Bravo-Leon, A., Morikawa, Y., Kawahara, M., Mayo, M.J., 2002. Fracture toughness of nanocrystalline tetragonal zirconia with low yttria content. Acta Mater. 50, 4555–4562. Bull, S.J., Chalker, P.R., Ayres, C.F., Rickerby, D.S., 1991. The influence of titanium interlayers on the adhesion of titanium nitride coatings obtained by plasma-assisted chemical vapordeposition. Mater. Sci. Eng. A 139, 71–78. Bull, S.J., 2011. Analysis methods and size effects in the indentation fracture toughness assessment of very thin oxide coatings on glass. Comptes Rendus Mecanique. 339, 518–531. Buser, D., Janner, S.F.M., Wittneben, J.G., Bragger, U., Ramseier, C.A., Salvi, G.E., 2012. 10-year survival and success rates of 511 titanium implants with a sandblasted and acid-etched surface: a retrospective study in 303 partially edentulous patients. Clin. Implant Dent. Relat. Res. 14, 839–851. Cao, Y.F., Allameh, S., Nankivil, D., Sathiaraj, T.S., Otiti, T., Soboyejo, W., 2008. Nanoindentation measurements of the mechanical properties of polycrystalline Au and Ag thin films on silicon substrates: effects of grain size and film thickness (vol 427, pg 232, 2006). Mater. Sci. Eng. A 494, 466 –466. Chen, J., Bull, S.J., 2006. Assessment of the toughness of thin coatings using nanoindentation under displacement control. Thin Solid Films. 494, 1–7. Chen, J., Bull, S.J., 2007. Indentation fracture and toughness assessment for thin optical coatings on glass. J. Phys. D-Appl. Phys. 40, 5401–5417. Chen, J.J., Bull, S.J., 2011. Approaches to investigate delamination and interfacial toughness in coated systems: an overview. J. Phys. D-Appl. Phys., 44. Chen, J.J., 2012. Indentation-based methods to assess fracture toughness for thin coatings. J. Phys. D—Appl. Phys., 45.
Cheng, C.C., Erdemir, A., Fenske, G.R., 1989. Correlation of interface structure with adhesive strength of ion-plated TiN hard coatings. Surf. Coat. Technol. 39–40 (Part 1), 365–376. Cheng, Y.T., Li, Z.Y., Cheng, C.M., 2002. Scaling relationships for indentation measurements. Philos. Mag. A 82, 1821–1829. Christel, P., Meunier, A., Heller, M., Torre, J.P., Peille, C.N., 1989. Mechanical properties and short-term in-vivo evaluation of yttrium-oxide-partially-stabilized zirconia. J. Biomed. Mater. Res. 23, 45–61. Cook, R.F., Pharr, G.M., 1990. Direct observation and analysis of indentation cracking in glasses and ceramics. J. Am. Ceram. Soc. 73, 787–817. Dierens, M., Vandeweghe, S., Kisch, J., Nilner, K., De Bruyn, H., 2012. Long-term follow-up of turned single implants placed in periodontally healthy patients after 16–22 years: radiographic and peri-implant outcome. Clin. Oral Implants Res. 23, 197–204. Doerner, M.F., Nix, W.D., 1986. A method for interpreting the data from depth-sensing indentation instruments. J. Mater. Res. 1, 601–609. Gerberich, W.W., Cordill, M.J., 2006. Physics of adhesion. Rep. Prog. Phys. 69, 2157–2203. Giannuzzi, L.A., Kempshall, B., Schwarz, S., Lomness, J., Prenitzer, B., Stevie, F., 2005. FIB Lift-out Specimen Preparation Techniques (Introduction to Focused Ion Beams). Springer201–228. Harding, D.S., Oliver, W.C., Pharr, G.M., 1995. Cracking during nanoindentation and its use in the measurement of fracture toughness. Thin Films: Stresses Mech. Prop. V 356, 663–668. Joss-Vassalli, I., Grebenstein, C., Topouzelis, N., Sculean, A., Katsaros, C., 2010. Orthodontic therapy and gingival recession: a systematic review. Orthodont. Craniofacial Res. 13, 127–141. Jung, R.E., Sailer, I., Ha¨mmerle, C.H., Attin, T., Schmidlin, P., 2007. In vitro color changes of soft tissues caused by restorative materials. Int. J. Periodont. Restor. Dent. 27, 251–257. Lawn, B.R., Evans, A.G., Marshall, D.B., 1980. Elastic/plastic indentation damage in ceramics: the median/radial crack system. J. Am. Ceram. Soc. 63, 574–581. Lawn, B.R., Howes, V.R., 1981. Elastic recovery at hardness indentations. J. Mater. Sci. 16, 2745–2752. Lu, M.Y., Xie, H.T., Huang, H., 2013. Characterisation of interfacial adhesion of thin film/substrate systems using indentationinduced delamination: a focused review. Prog. Surf. Treat. Ii. 533, 201–222. Mesarovic, S.D., Fleck, N.A., 1999. Spherical indentation of elastic– plastic solids. Proc. R. Soc. London Ser. A 455, 2707–2728. Oliver, W.C., Pharr, G.M., 1992. An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments. J. Mater. Res.. 7, 1564–1583. Pecnik, C.M., Roos, M., Muff, D., Spolenak, R., Sailer, I., 2014a. In vitro color evaluation of esthetic coatings for metallic dental implants and implant prosthetic appliances. Clin. Oral Impl. Res.http: //dxdoi.org/10.1111/clr.12443 First published online: 15.07.2014. Pecnik, C.M., Muff, D., Spolenak, R., Sailer, I., 2014b. Color evaluation of a dielectric mirror coating using porcine tissue and prosthetic gingival material: a comparison of two models. Clin. Oral Impl. Res.. Pelleg, J., 2012. Mechanical Properties of Materials. Springer. Petzow, G., 1999. Metallographic Etching: Techniques for Metallography, Ceramography, Plastography, second ed. ASM International. Pharr, G.M., 1998. Measurement of mechanical properties by ultra-low load indentation. Mater. Sci. Eng. A 253, 151–159. Rickerby, D.S., Burnett, P.J., 1988. Correlation of process and system parameters with structure and properties of physically vapourdeposited hard coatings. Thin Solid Films. 157, 195–222. Ro¨sler, J., Harders, H., Ba¨ker, M., 2007. Mechanical Behaviour of Engineering Materials: Metals, Ceramics, Polymers, and Composites. Springer, Berlin; New York.
journal of the mechanical behavior of biomedical materials 47 (2015) 1–11
Saha, R., Nix, W.D., 2002. Effects of the substrate on the determination of thin film mechanical properties by nanoindentation. Acta Mater. 50, 23–38. Schiffmann, K.I., 2011. Determination of fracture toughness of bulk materials and thin films by nanoindentation: comparison of different models. Philos. Mag. A 91, 1163–1178.
11
van Brakel, R., Noordmans, H.J., Frenken, J., de Roode, R., de Wit, G.C., Cune, M.S., 2011. The effect of zirconia and titanium implant abutments on light reflection of the supporting soft tissues. Clin. Oral Impl. Res. 22, 1172–1178.
