Evaluating interface strength of calcium phosphate sol-gel-derived thin films to Ti6Al4 V substrate

Lu Gan ^{a, b}, Jian Wang ^{a, c} and Robert M. Pilliar <sup>, , a, b, c

a</sup> Institute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, ON, Canada M5S 3E3
^b Department of Materials Science and Engineering, University of Toronto, Toronto, ON, Canada M5S 3E3
^c Faculty of Dentistry, University of Toronto, Toronto, ON, Canada M5S 3E3

Received 20 November 2003;  accepted 5 February 2004.  Available online 11 March 2004.
Biomaterials
Article in Press, Corrected Proof - Note to users

1. Abstract

The interface shear strength of Ca-P thin films applied to Ti6Al4 V substrates have been evaluated in this study using a substrate straining method—a shear lag model. The Ca-P films were synthesized using sol-gel methods from either an inorganic or organic precursor solution. Strong interface bonding was demonstrated for both film types. The films were identified as non-stoichiometric hydroxyapatite but with different Ca/P ratios. The Ca-P films were 1-1.5 
m thick and testing and analysis using the shear lag approach revealed a shear strength of approximately 347 and 280 MPa for Inorganic and Organic Route-formed films, respectively. Overall, the exceptional mechanical properties of Ca-P/Ti6Al4 V system along with the inherent advantages of sol-gel processing support continued studies to utilize this technology for bone-interfacing implant surface modification.

Author Keywords: Author Keywords: Calcium phosphate coating; Sol-gel; Ti6Al4 V; Interfacial strength; Shear lag strain test

2. Article Outline

1. Introduction

2. Materials and methods

2.1. Sample preparation and film characterization

2.2. Interfacial shear strength testing

3. Results

3.1. Film characterization

3.2. Interfacial shear strength determination

4. Discussion

4.1. Relevance of interfacial shear strength values

4.2. Adhesion bonding mechanisms

5. Conclusions

Acknowledgements

References

3. 1. Introduction

A number of techniques have been reported for synthesizing calcium phosphate (Ca-P)/hydroxyapatite (HA) coatings on metallic substrates, primarily for the purpose of enhancing osteoconductivity of bone-interfacing implants. Sol-gel thin film processing has been shown to be a relatively simple procedure utilizing low temperature anneals to form Ca-P thin films of uniform structure and composition. Recent studies in our laboratory have demonstrated that films can be formed using either inorganic or organic precursor solutions with both approaches resulting in non-stoichiometric hydroxyapatite [1]. Differences in Ca/P ratio and film ultrastructure were noted for films formed by one approach or the other. Enhanced osteoconductivity of Inorganic Route-formed films was demonstrated through in vivo studies of implants placed in femoral condyle sites of New Zealand White rabbits [2]. However, a necessary requirement for successful use of such coatings with load-bearing implants is the assurance of formation and maintenance for as long as required of a strong bond at the Ca-P film-substrate interface. Few studies have investigated Ca-P film-to-substrate adhesion, even though acceptable implant performance is more or less contingent on good bonding at that interface.

Substrate straining tests have been used to investigate the mechanical behavior and bonding of thin brittle films with ductile substrate materials [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 and 16]. In these tests, tensile force is applied to the more ductile substrate causing transverse crack development in the more brittle coating overlayer. As the tensile force is increased, crack frequency increases with a steady-state crack density and inter-crack spacing developing. The interfacial shear strength,
_max, has been related to the steady-state inter-crack spacing by various relationships of the general form

max= f/k ,

(1)

where
is a measure of the steady-state crack spacing (either minimum, mean or maximum spacing),
is the film thickness,
_f is the tensile fracture strength of the film, and k is a dimensionless constant [3, 6, 7 and 14].

In the present study, we applied a substrate straining test and shear lag analysis method for evaluating the interface shear strength of Ti6Al4 V samples with thin sol-gel-formed Ca-P films made using either inorganic or organic precursor solutions. We have used this test method previously for determining the interfacial shear strength of sol-gel-derived ZrO₂ films on Ti6Al4 V [12] and stainless steel [14 and 15] substrates. As reported elsewhere [1 and 2], we are investigating these HA films as coatings to increase the rate of bone ingrowth into porous-surfaced Ti6Al4 V implants.

4. 2. Materials and methods

2.1. Sample preparation and film characterization

Mill-annealed Ti6Al4 V samples of two different forms were used in these studies. Polished disc samples, (600-grit finish), of 12.7 mm diameter×1.5 mm thickness, were prepared for surface characterization of the sol-gel-formed Ca-P films. In addition, standard wedge-gripped tensile specimens (ASTM E8) with a 1 
m surface finish in the gage length region were used to determine the interfacial shear strength using the substrate straining method described below. The sol-gel layer was applied by dip coating. In preparation for sol-gel coating, substrates were degreased using acetone and then washed with detergent followed by copious rinsing with distilled water. Finally, the samples were given a nitric acid passivation treatment prior to the coating operation.

The procedures used for forming the Ca-P films using the Inorganic or Organic precursor solutions have been described in detail elsewhere [1]. Briefly, for Inorganic Route-formed films, calcium nitrate tetrahydrate and ammonium dihydrogen phosphate were used as calcium and phosphorous precursors, and the molar ratio of Ca/P in these starting reactants was set equal to 1.67 [17]. Five separate film layers were applied using a substrate withdrawal rate of 30 cm/min and relatively low temperature anneals (~210°C) between dippings in order to develop a final film thickness ~1 
m. A final anneal at 500°C for 10 min in air was used to consolidate the Ca-P film followed by sample furnace cooling to room temperature. For the Organic Route-formed film, calcium nitrate tetrahydrate and triethyl phosphite were selected as the calcium and phosphorous precursors [18 and 19]. The calcium nitrate tetrahydrate was dissolved in absolute anhydrous ethanol that was then added drop-wise into the hydrolyzed phosphorus solution, again at a Ca/P ratio equal to 1.67. This clear solution was aged in a 40°C water bath for 4 days. A diluted sol from the mother solution was further aged for 2 days at room temperature to form the film for substrate coating. A single dipping treatment was used with the Organic Route solution. The substrates were withdrawn from the solution at a rate of 20 cm/min and the resulting coating annealed at 500°C for 20 min in air and then furnace cooled to room temperature.

Thin Film X-ray Diffraction (TFXRD) was used to determine the crystal structure of the films. Scans were made over the range 2
=20-60° in increments of 0.01°. Diffuse-Reflectance Infrared Fourier Transform (DRIFT) Spectroscopy was used to determine the nature of functional groups at the film surface. All IR spectra were obtained at 4 cm⁻¹ resolution averaging 240 scans in a reflective mode with a Nicolet Nexus 670 FT-IR system. The surface composition of the films was determined using X-ray Photoelectron Spectroscopy (XPS) with Mg K
X-ray radiation. Film thickness was determined by examination of cross-sections of coated samples using scanning electron microscopy (SEM).

2.2. Interfacial shear strength testing

Given the testing constraints imposed by the sol-gel film thickness (~1 
m), the shear lag strain method developed by AGRAWAL and RAJ [6 and 7] was used to assess the film-substrate interface shear strength.

Testing was conducted using an Instron 8501 servohydraulic mechanical testing machine operated under strain control with an extensometer attached to the gage length region of the test specimens (Instron model 2620-830) to monitor precise strain levels. The interfacial shear strength
_max was determined using a simplified equality for a film of thickness
[7],

max= f/1.5 ,

(2)

where
is the average steady-state crack spacing and
_f is the tensile strength of the film.
_f was determined experimentally by measuring the strain,
_f, at which cracks were first detected to form and multiplying by the film modulus.

To capture
_f, samples were strained sequentially in 0.2% increments, followed by SEM examination of random areas within the central gage region in order to detect cracks that may have formed. Larger strain increments were only employed to reach the steady-state crack density and spacing. A systematic collection of crack spacing data for each sample and strain level was performed using SCION^® image analysis software. Here, crack spacing was defined as the distance between two successive transverse cracks. Several random area micrographs within the central gage region at 3000× magnification were taken and evaluated. Routine statistical analysis of the crack spacing data provided an average steady-state crack spacing that was used to calculate the interfacial shear strength.

The film elastic modulus, also required for determination of the film tensile strength,
_f, was determined using a nanoindentation deformation technique. Nanoindentation testing was performed using a TriboScope^® nanomechanical testing system (Hysitron, Inc., USA) using the method described below and the data analysis method developed by OLIVER and PHARR [20].

A cube-corner indenter was used to apply low and ultra-low-load indentations in selected locations of the Ca-P-coated samples. A highly polished mill-annealed Ti6Al4 V disc (1 
m diamond finish, R_a=0.030±0.008 
m) was used as a standard sample. The modulus values of this standard sample were compared with literature values to verify the validity and accuracy of this nanoindentation technique.

The indentation tests were conducted using a load-unload test mode. The displacement of the indenter was continuously monitored and a load-time history of the indentation recorded. After reaching the maximum load, the indenter was held for 30 s before unloading. After unloading 90% of the peak load, the indenter was held for 30 s to establish the rate of thermal drift in the machine and in the specimen for purposes of data correction. Five indentations were performed in each local testing area. The load-displacement curves were recorded and the top unloading curves were used to calculate the reduced modulus, E_r.

(3)

where S is the slope of the initial portion of the unloading curve and A is the indenter project area of contact [20]. The modulus of the Ca-P film can then be calculated as follows:

(4)

where E_m and
_m are the modulus and Poisson's ratio of the indented Ca-P film, respectively, and E_i and
_i are those of the diamond indenter, given as E_i=1141 GPa and
_i=0.07. The value used for the Poisson's ratio of the Ca-P film was equal to 0.28 [21].

5. 3. Results

3.1. Film characterization

SEM examination of the disc specimens and the tensile specimens prepared with Inorganic Route-formed Ca-P films showed that the substrate was completely covered by a rough film (R_a=0.272±0.002
m). SEM examination of a polished cross-section section of the film indicated a mean film thickness equal to 943±25 nm. The Organic Route-formed films generally were not as rough and reflected the surface quality of the underlying substrate with the presence of occasional pinhole-like defects likely due to microscopic-sized debris that may have settled on the sol or substrate surfaces prior to or during the dip-coating process. No other film cracking was observed, except for gross cracking found at substrate geometric discontinuities (i.e. at sample edges).

The results of the crystal structure and the chemical compositions of both Ca-P films have been reported elsewhere [1]. Briefly, TFXRD spectra of both samples indicated the formation of an HA structure characterized by broad diffraction peaks indicating either a high dislocation density giving rise to lattice distortion or the presence of a very fine-grained, nanocrystalline structure, or a combination of both these factors. DRIFT spectra indicated both films contained hydroxyl, carbonate and phosphate groups. XPS analysis of the samples showed that the Ca/P ratio was about 1.46 for the Inorganic Route-formed film, while it was about 2.10 for the Organic Route-formed film indicating the formation of non-stoichiometric carbonated calcium hydroxyapatite by both routes.

3.2. Interfacial shear strength determination

Despite its somewhat irregular surface topography, the tensile specimens prepared using the Inorganic Route coating approach displayed the development of clearly discernible transverse cracks that approached a steady-state crack density and inter-crack spacing with increasing applied strain. A typical sequence of crack pattern development towards this steady-state is shown in Fig. 1. No cracking was noted below the yield strain of the Ti6Al4 V substrate (~0.8%). At a substrate strain level of 1.8%, fine non-continuous transverse cracks could be detected in all areas examined. With increasing strain, a regular transverse crack pattern developed in the films until at higher strains (~7%) little apparent change in crack spacing could be noted, although isolated film delamination did become more common. The steady-state crack spacing was equal to 3.28±0.02 
m. A typical crack spacing distribution at steady-state (
=6%) as a function of plastic strain is illustrated in Fig. 2. The inter-crack spacing measurements were close to a normal distribution. This steady-state distribution was achieved rather quickly, as emphasized in a plot of crack density (the inverse of transverse crack spacing
) versus strain in Fig. 3.

(30K)

Fig. 1. Scanning electron micrographs of typical transverse crack pattern at (a)
=1.8%, (b)
=6%. The Ca-P films were formed using the Inorganic Route.

(12K)

Fig. 2. Typical crack spacing distribution for Inorganic Route-formed films at steady-state strain level,
=6%. This distribution is close to normal distribution.

(4K)

Fig. 3. Crack density as a function of strain for Inorganic Route-formed films indicating crack density, the reciprocal of crack spacing, approaches steady state after
=6%.

For the Organic Route-formed films, similar trends were observed. Transverse cracks were first detected at a substrate strain level of 0.6%, which was lower than the yield strain of Ti6Al4 V substrate (~0.8%). Steady-state crack spacing was achieved rapidly at strain levels ~1.5%, (Fig. 4). The steady-state inter-crack spacing at
~2.5% was again close to a normal distribution (Fig. 5) with a mean and standard deviation of 3.05±0.02 
m. The crack density versus strain is shown in Fig. 6.

(27K)

Fig. 4. Scanning electron micrographs of typical transverse crack pattern at (a)
=0.6%, and (b)
=2.5%. The Ca-P films were formed using the Organic Route.

(12K)

Fig. 5. Typical crack spacing distribution for Organic Route-formed films at steady-state strain level
=2.5%. This distribution is close to normal distribution.

(4K)

Fig. 6. Crack density as a function of strain for Organic Route-formed films indicating crack density, the reciprocal of crack spacing, approaches steady state after
=1.5%.

The modulus calculated from the nanoindentation test for the Inorganic Route-formed films was equal to approximately 32.1 GPa while for the Organic Route-formed films it was determined to be 45.4 GPa. This difference in modulus was attributed to a significant difference in film density as reported in the related paper by Gan et al. [1]. Substituting all parameters into the shear lag equation, the maximum interfacial shear strengths of the Inorganic and Organic Route-formed Ca-P films were equal to approximately 347 and 280 MPa, respectively (Table 1). It should be stressed that the accuracy of the modulus measured by the nanoindentation test plays an important role in determining the maximum interfacial shear strength in the substrate straining test. The average modulus of the mill-annealed Ti6Al4 V sample was determined by this method to be 116±3 GPa. This value agrees well with the literature value for the elastic modulus determined using conventional test samples and methods (E=113.8 GPa) [22 and 23]. The modulus for HA reported in literature ranges from 40-90 GPa, depending on material porosity [24 and 25]. The modulus values of our sol-gel-formed films determined using the nanoindentation testing, therefore, appear reasonable.

Table 1. Interfacial shear adhesion strength and related film parameters

6. 4. Discussion

4.1. Relevance of interfacial shear strength values

The interfacial shear strengths for the both Inorganic and Organic Route-formed sol-gel HA films are roughly an order of magnitude greater than the shear adhesion strengths reported for plasma-sprayed HA coatings [26 and 27]. In contrast, sol-gel derived Ca-P films at least initially appear to have an interfacial shear strength that is of the same order of magnitude as submicron HA films produced by a pulsed laser deposition process. The interfacial bond strengths of these films were estimated using scratch testing and failure pattern analysis and shown to be in excess of the yield strength of the Ti alloy (~930 MPa) [28]. From an adhesion standpoint, then, sol-gel derived Ca-P films appear quite promising for load-bearing metal implant application.

Determination of interface shear strength using the shear lag analysis assumes a limited film thickness and, hence, low transverse residual stress within the film. Excessive transverse stresses would result in film delamination. In our studies, film delamination was not observed for applied strains corresponding to just achieving steady-state crack density although at higher substrate strains (i.e. >6% for the Inorganic Route-formed films and >3% for the Organic Route-formed films) delamination was observed.

The derivation of
_max is based on the assumption of a coating free of residual stress [6 and 7]. Hence, for a coating with residual stress,
_R, the effective stress,
_e, that initiates the fragmentation of the coating must be adjusted to include the contribution from residual stress. Accordingly,
_e=
_f+
_R, where
_f=E
_f is the apparent fracture strength of the coating [14]. Residual strain results in both peak shifts and peak broadening of diffraction peaks in TFXRD data. However, peak broadening is also related to crystal size with fine crystal size also causing peak broadening. Differentiating between strain and crystal size changes can be difficult using TFXRD data alone. The peak shifts (compared to International Center for Diffraction Data file 9-0432 for HA powders) observed with the Inorganic and Organic Route-formed Ca-P films suggested that residual stress was present in these coatings [1]. However, we were unable to determine the magnitude of any residual stress so this correction to
_e was not included in our analysis. This factor may contribute to an underestimation of the true value of residual stress within the film, although this may be countered by other factors as discussed below.

There are a number of possible causes of such residual film stresses. For sol-gel derived films, it is commonly accepted that film shrinkage driven by capillary stresses occurring during drying and annealing of the films contributes to a film thickness threshold (~0.5 
m for a single coating) above which film cracking routinely occurs [29]. For well-adhering films in which in-plane shrinkage is inhibited, the resulting residual tensile stresses (or strains) in the un-cracked films should promote crack development at lower applied strains during shear lag testing and result in a reduction of the measured failure strength
_f and failure strain
_f. The coefficient of thermal expansion (CTE) of HA is higher than that of Ti6Al4 V [30, 31 and 32], (see Table 2). Therefore, residual thermal stresses due to a mismatch between the thermal coefficient of expansion of the Ca-P coating and the Ti6Al4 V substrate [33, 34 and 35] will contribute to residual tensile strains and stresses in the film. Residual strain, can also be due to atom or molecule substitution within the HA lattice. Carbonate groups did appear to substitute within the HA lattice as evidenced by the FTIR spectra [1] of the films with both Inorganic and Organic Route-formed HA films forming Type B carbonated HA.

Table 2. Some mechanical properties of Ti6Al4 V and HA (30-32)

In calculating
_max using Eq. (2),
_f introduces the greatest possible source of error since it is dependent on the detection of the strain causing transverse crack initiation and on an accurate measurement of the film modulus using the nanoindentation loading method. The results of the analysis of the load-deflection curves from the nanoindentation testing indicated a higher E for the Organic Route-formed film (Table 1). This seemed reasonable in view of the observed denser structure of this film compared with the Inorganic Route-formed film [1]. However, the much higher apparent value of
_f for the Inorganic Route-formed film (578 vs. 272 MPa) can be questioned. Determination of the critical strain for crack initiation was achieved by examination of sample surfaces using SEM after different strain increments. This was dependent on the detection of distinct cracks in regions of the film surface viewed at 3000× magnification. Development and detection of significant cracks in the dense continuous Organic Route-formed film was easier than in the lower density Inorganic Route-formed film. It can be argued that the somewhat irregular (`orange-peel') texture of the latter film surface and its fine nano-scale structure may have inhibited the growth of cracks to a detectable length and made detection of initial cracks more difficult. Thus, we may have over-estimated
_f for this film resulting in an apparent higher
_max. In addition to the possible effects due to residual stresses in the film, the determined
_max should be considered as order of magnitude estimates only. Whether true differences exist in terms of interface bonding to Ti6Al4 V substrates between Organic and Inorganic Route-formed HA films requires further study. Nevertheless, the high values for both film types suggests their potential usefulness over other methods for modifying Ti6Al4 V implants to enhance osteoconductivity.

4.2. Adhesion bonding mechanisms

The high interface shear strengths determined for both film types (i.e. Inorganic and Organic Route-formed) suggest strong primary chemical bond development at the film-substrate interface. High-resolution examination of the interface region using transmission electron microscopy and associated selected area electron diffraction have indicated the formation of an interface reaction product (calcium titanium oxide-CaTi₂O₅) for the Inorganic Route-formed film [1]. While a similar product was not detected with the Organic Route-formed film, this may have been due to the shorter cumulative time at higher temperatures for the film formed by this method (i.e. 500^oC for 20 min only vs. five pre-anneals at 210°C + the 500°C, 10 min final anneal). Differences due to pre-cursor film reactivity (density, chemistry, structure) may also have caused this difference. Another model had been proposed previously to explain high interfacial shear strengths for ZrO₂ sol-gel-formed films adhering to Ti6Al4 V substrates [12]. This involved interactions at the interface between alkoxide molecules in the film and free hydroxyl groups at the metal oxide surface (the thin TiO₂ oxide that forms spontaneously on the Ti alloy) via a condensation reaction similar to that occurring between alkoxide molecules. A similar condensation-type reaction may occur between alkoxide HA precursors used in the Organic Route process and free hydroxyl groups associated with the passivated TiO₂ layer. It is interesting to note that the interface shear strength for the Organic Route-formed Ca-P film was similar to the value determined for the ZrO₂ sol-gel-formed film on Ti6Al4 V in that previous study [12]. Both films were dense and were formed using similar annealing conditions. While direct evidence for such bonding was not found, a condensation reaction at this interface at the 500°C annealing temperature may explain the observed high interfacial shear strength.

7. 5. Conclusions

The substrate straining method and application of the shear lag approach proved useful in assessing interface shear strength of sol-gel derived Ca-P thin films applied to Ti6Al4 V substrates. The relatively high strength values determined for both Inorganic and Organic Route-formed films are believed to be a result of primary chemical bonding that result during the heat treatment used to form the HA films. Considering the simplicity of the sol-gel dip-coating method and the good interfacial strength of the resulting HA films, sol-gel processing offers a promising approach for modifying the surface of load-bearing implants intended for bone-interfacing applications.

8. Acknowledgements

The authors would like to thank Dr. David Bahr and Ms. Megan Cordill at Washington State University for undertaking the nanoindentation testing and Dr. Christopher Yip and Mr. Patrick Yang at the University of Toronto for the FTIR testing. Funding from Natural Sciences & Engineering Research Council of Canada is gratefully acknowledged.

9. References

1. Gan L, Pilliar RM. Calcium phosphate sol-gel-derived thin films on porous-surfaced implants for enhanced osteoconductivity: I. Synthesis and characterization, Biomaterials, in press.

2. A. Tache, L. Gan, R.M. Pilliar and D. Deporter, Effect of surface chemistry on the rate of osseointegration of sintered porous-surfaced Ti-6Al-4 V implants. Int J Oral Max Impl 19 (2004), pp. 19-29. Abstract-MEDLINE  

3. Wojciechowski PH, Mendolia MS. Fracture and cracking phenomena in thin films adhering to high-elongation substrates. In: Francombe MH, Vossen JL, editors. Physics of thin films 16. Boston: Academic Press; 1992. p. 271-340.

4. Q. Guo, H. Osaki and L.M. Keer, Measurement of the intrinsic bond strength of brittle thin-films on flexible substrates. J Appl Phys 68 (1990), pp. 1649-1654. Abstract-INSPEC   | Full Text via CrossRef

5. P.H. Duchatelard, G. Baud, J.P. Besse and M. Jacquet, Alumina coatings on PMMA—Optimization of adherence. Thin Solid Films 250 (1994), pp. 142-150. Abstract

6. D.C. Agrawal and R. Raj, Measurement of the ultimate shear strength of a metal-ceramic interface. Acta Metall 37 (1989), pp. 1265-1270. Abstract

7. D.C. Agrawal and R. Raj, Ultimate shear strengths of copper-silica and nickel-silica interfaces. Mater Sci Eng A126 (1990), pp. 125-131. Abstract

8. L. Chandra, M. Allen, R. Butter, N. Rushton, A.H. Lettington and T.W. Clyne, The effect of exposure to biological fluids on the spallation resistance of diamond-like carbon coatings on metallic substrates. J Mater Sci-Mater Med 6 (1995), pp. 581-589. Abstract-EMBASE | Abstract-Compendex  

9. L. Chandra and T.W. Clyne, Characterization of the strength and adhesion of diamond films on metallic substrates using a substrate plastic straining technique. Diam Relat Mater 3 (1994), pp. 791-798. Abstract

10. P.M. Ramsey, H.W. Chandleer and T.F. Page, Bending tests to estimate the through thickness strength and interfacial shear strength in coated systems. Thin Solid Films 201 (1991), pp. 81-89. Abstract

11. J.H. Jeong and D. Kwon, Evaluation of the adhesion strength in DLC film-coated systems using the film-cracking technique. J Adhes Sci Technol 12 (1998), pp. 29-46. Abstract-Compendex  

12. M.J. Filiaggi, R.M. Pilliar and D. Abdulla, Evaluating sol-gel ceramic thin films for metal implant applications. II. Adhesion and fatigue properties of zirconia films on Ti6Al4 V. J Biomed Mater Res 33 (1996), pp. 239-256. Abstract-MEDLINE | Abstract-EMBASE | Abstract-Compendex   | Full Text via CrossRef

13. P.B. Kirk and R.M. Pilliar, The deformation response of sol-gel-derived zirconia thin films on 316L stainless steel substrates using a substrate straining test. J Mater Sci 34 (1999), pp. 3967-3975. Abstract-Compendex | Abstract-INSPEC   | Full Text via CrossRef

14. B.F. Chen, J. Hwang, I.F. Chen, G.P. Yu and J.H. Huang, A tensile-film-cracking model for evaluating interfacial shear strength of elastic film on ductile substrate. Surf Coat Tech 126 (2000), pp. 91-95. SummaryPlus | Full Text + Links | PDF (540 K)

15. B.F. Chen, J. Hwang, G.P. Yu and J.H. Huang, In situ observation of the cracking behavior of TiN coating on 304 stainless steel subjected to tensile strain. Thin Solid Films 352 (1999), pp. 173-178. SummaryPlus | Full Text + Links | PDF (160 K)

16. M.S. Hu and A.G. Evans, The cracking and decohesion of thin-films on ductile substrates. Acta Metall Mater 37 (1989), pp. 917-925. Abstract

17. Q. Qiu, P. Vincent, B. Lowenberg, M. Sayer and J.E. Davies, Bone growth on sol-gel calcium phosphate thin films in vitro. Cell Mater 3 (1993), pp. 351-360.

18. D.M. Liu, T. Troczynski and D. Hakimi, Effect of hydrolysis on the phase evolution of water-based sol-gel hydroxyapatite and its application to bioactive coatings. J Mater Sci-Mater Med 13 (2002), pp. 657-665. Abstract-EMBASE | Abstract-Compendex   | Full Text via CrossRef

19. D.M. Liu, Q.Z. Yang and T. Troczynski, Sol-gel hydroxyapatite coatings on stainless steel substrates. Biomaterials 23 (2002), pp. 691-698. SummaryPlus | Full Text + Links | PDF (401 K)

20. W.C. Oliver and G.M. Pharr, An improved technique for determining hardness and elastic-modulus using load and displacement sensing indentation experiments. J Mater Res 7 (1992), pp. 1564-1583. Abstract-INSPEC  

21. L.H. Guo, X. Guo, Y. Leng, J.C.Y. Cheng and X.D. Zhang, Nanoindentation study of interfaces between calcium phosphate and bone in an animal spinal fusion model. J Biomed Mater Res 54 (2001), pp. 554-559. Abstract-MEDLINE | Abstract-EMBASE   | Full Text via CrossRef

22. Bever MB. Encyclopedia of materials science and engineering. New York: Pergamon Press; 1986. 1059 p.

23. Callister WD. Materials science and engineering, an introduction. New York: Wiley; 2000. 793 p.

24. H. Aoki, Science and medical applications of hydroxyapatite. , Takayama Press System Centre Company, Tokyo (1991).

25. H.H. Moroi, K. Okimoto, R. Moroi and Y. Terada, Numeric approach to the biomechanical analysis of thermal effects in coated implants. Int J Prosthodont 6 (1993), pp. 564-572. Abstract-MEDLINE  

26. Kay JF, Jarcho M, Logan G, Liu ST. The structure and properties of hydroxylapatite coatings on metal. In: Transactions of 12th Annual Meeting Society for Biomaterials 1986, p. 13.

27. F. Bross, A. Cigada, R. Chiesa, L. Paracchini and C. Consonni, Post-deposition treatment effects on hydroxyapatite vacuum plasma spray coatings. J Mater Sci-Mater Med 5 (1994), pp. 855-857.

28. C.M. Cotell, D.B. Chrisey, K.S. Grabowski, J.A. Sprague and C.R. Gossett, Pulsed laser deposition of hydroxyapatite thin films on Ti6Al4 V. J Appl Biomater 3 (1992), pp. 87-93.

29. C.J. Brinker and G.W. Scherer. Sol-gel science: the physics and chemistry of sol-gel processing, Academic Press, Boston (1990).

30. S. Nakamura, R. Otsuka, H. Aoki, M. Akao, N. Miura and T. Yamamoto, Thermal expansion of hydroxyapatite-
-tricalcium phosphate ceramics. Thermochim Acta 165 (1990), pp. 57-72. Abstract

31. S. Maruno, K. Hayashi, Y. Sumi, Y.F. Wang and H. Iwata. In: T. Yamamuro, L.L. Hench and J. Wilson, Editors, CRC Handbook of bioactive ceramics, CRC Press, Boca Raton (1991), pp. 187-193.

32. K. Akagi, Y. Okamoto, T. Matsuura and T. Horibe, Properties of test metal ceramic titanium alloys. J Prosthet Dent 68 (1992), pp. 462-467. Abstract-MEDLINE  

33. J.A. Jansen, J.G.C. Wolke, S. Swann, J.P.C.M. Van Der Waerden and K. de Groot, Application of magnetron sputtering for producing ceramic coatings on implant materials. Clin Oral Impl Res 4 (1993), pp. 28-34. Abstract-MEDLINE   | Full Text via CrossRef

34. K. Vandijk, H.G. Schaeken, J.C.G. Wolke, C.H.M. Maree, F.H.P.M. Habraken, J. Verhoeven and J.A. Jansen, Influence of discharge power level on the properties of hydroxyapatite films deposited on Ti6Al4 V with RF magnetron sputtering. J Biomed Mater Res 29 (1995), pp. 269-276. Abstract-MEDLINE | Abstract-Compendex  

35. S.J. Ding, C.P. Ju and J.H.C. Lin, Immersion behavior of RF magnetron-assisted sputtered hydroxyapatite/titanium coatings in simulated body fluid. J Biomed Mater Res 47 (1999), pp. 551-563. Abstract-MEDLINE | Abstract-EMBASE | Abstract-Compendex   | Full Text via CrossRef

Corresponding author. University of Toronto, , Toronto, , Canada. Tel.: [phone]; fax: [phone]

---