- Email: [email protected]
mixing deposition
Nuclear Instruments and Methods in Physics Research B 179 (2001) 364±372
www.elsevier.com/locate/nimb
Fabrication and characterization of graded calcium phosphate coatings produced by ion beam sputtering/mixing deposition C.X. Wang a
c
a,b,*
, Z.Q. Chen a, L.M. Guan a, M. Wang b, Z.Y. Liu c, P.L. Wang c
Department of Dental Materials, College of Stomatology, West China University of Medical Sciences, Chengdu 610041, Sichuan, China b School of Mechanical and Production Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore Laboratory of Radiation Physics and Technology, Institute of Nuclear Science and Technology, Sichuan University, Chengdu 610064, Sichuan, China Received 2 January 2001; received in revised form 23 February 2001
Abstract Ion beam sputtering/mixing deposition was used to produce thin calcium phosphate coatings on titanium substrate from the hydroxyapatite target. It was found that as-deposited coatings were amorphous. No distinct absorption band of the hydroxyl group was observed in FTIR spectra of the coatings but new absorption bands were present for CO23 , which was brought about during the deposition process. Scanning electron microscopy revealed that the deposited coatings had a uniform and dense structure. The calcium to phosphorous ratio of these coatings varied between 2.0 and 8.0. Analyses of XPS data revealed that the coating could be divided into four distinctive zones, and a graded structure was achieved in the as-received coating. Scratch tests showed that the coatings adhered well to the substrate. Ó 2001 Elsevier Science B.V. All rights reserved. Keywords: Ion beam sputtering deposition; Hydroxyapatite; Calcium phosphate coating; Titanium substrate; Characterization
1. Introduction Hydroxyapatite (HA) is a biocompatible and bioactive material. It is well known for its osteoconductivity [1,2]. However, the brittle nature of
*
Corresponding author. Tel.: +65-790-4963; fax: +65-7911859. E-mail address: [email protected] (C.X. Wang).
HA ceramic limits its clinical use in load-bearing situations such as total hip replacement. HA, nevertheless, has been used as a coating for metal implants, because it can speed up bone healing adjacent to implants and form a strong chemical bond with bone [3,4]. Currently, commercially available coated metal implants are manufactured by using the plasma spray technique for depositing HA coatings. However, problems such as low bond strength
0168-583X/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 5 8 3 X ( 0 1 ) 0 0 5 8 5 - 7
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
between the coating and the substrate and nonuniformity across the thickness of the coating are often encountered with plasma sprayed coatings [5]. A variety of other coating techniques, such as electrochemical deposition [6], radio-frequency magnetron sputtering [7], excimer laser deposition [8], pulsed laser deposition [9] and dipping [10], have also been investigated for producing HA coatings on metallic substrates. Since the mid1970s, many surface modi®cation techniques that are based on ion implantation, such as ion beam assisted deposition (IBAD), ion beam deposition (IBD), ion beam mixing (IBM), and techniques that are based on plasma-assisted ion implantation such as plasma source ion implantation (PSII) and plasma immersion ion implantation (PIII), have been developed rapidly and are now widely used to modify the surface of materials including metals, ceramics and polymers [11]. Ion beam sputter deposition was investigated as a method for producing biocompatible ceramic coatings on metallic implants due to its advantages which include the production of thin coatings with high density and superior adhesion [12,13]. In this process, the ionized-argon gas was used to sputter atoms from a ceramic target. The sputtered atoms built up on the metallic substrate that was placed in the path of the sputtered material. Both argon ion beam and nitrogen ion beam can be used for the ion beam bombardment. It was found that grafting ±NH 2 amidogen radicals on Al2 O3 ceramic using ion implantation technique could improve the biocompatibility of this ceramic [14]. So if using nitrogen ion beam to bombard the Ca±P coatings, it is possible to further improve the bone bonding between the coating and bone to expedite the wound healing. Moreover, during the bombardment process, nitrogen ions could penetrate into the whole coating layer and bond with the titanium substrate to form Tix Ny , which is bene®cial to improve the properties of coatings. In this investigation, the ion beam sputtering/mixing deposition technique was used to produce thin calcium phosphate coatings on titanium substrate. The emphasis was placed on the structural and chemical characterizations of the coatings produced by nitrogen ion beam mixing in this paper.
365
2. Materials and methods A commercially available pure titanium substrate (plates of dimensions 20 mm 20 mm 1 mm) was mechanically polished and ultrasonically cleaned with acetone and alcohol. HA powder was synthesized using the wet method. HA powder disks, which were to be used as the ion beam sputtering target, were cold-pressed and sintered at 1150°C for 3 h. X-ray diraction patterns of HA disks matched that of the standard synthetic hydroxyapatite (PDFSM #9-0432), indicating that the disks had well crystallized HA with a microstructure of purely random grain orientation. The ion beam sputtering deposition system has been described previously [15]. It mainly consisted of one Kaufman ion source and one Freeman ion source, one target holder and one rotatable sample holder in the path of both ion beams. The deposition chamber was evacuated to a base pressure of 2:8±3:7 10 4 Pa. Prior to deposition, etching substrates with 800 eV and 40 mA cm 2 argon ions for 30 min was performed to clean the surface of titanium substrate. The energetic ion beam was produced by ionizing high purity argon gas (99.999% pure). After cleaning, the stage was rotated so that the substrates were placed in the path of the sputtered atoms. The sputtering deposited monolayer coatings were produced using an Ar beam of 900 eV and 20 mA cm 2 for 3 h, 1200 eV and 40 mA cm 2 for 90 min and 1500 eV and 60 mA cm 2 for 60 min, respectively. For the Ar beam sputtering and N beam mixing deposited Ca±P coatings, after sputtering the HA target with 1200 eV and 40 mA cm 2 Ar beam for 1.5 h, the N beam with 60 KeV was used to homogenize the coating and its dosage was 2:0 1016 ions cm 2 . The thickness of all the coatings is about 300 nm. X-ray diraction (XRD) was employed to analyze the structure of as-sputtered coatings. A Rigaiku D/max-sA X-ray diractometer with Cu Ka radiation at 40 KeV and 50±80 mA was used. Transmission electron microscopy (TEM) was also used to examine the microstructure of as-deposited coatings. Because of the brittle nature of thin Ca±P coatings, it was very dicult to prepare suitable specimens for TEM observation directly from coatings with titanium substrate by using an ion
366
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
beam thinning apparatus. An alternative method was used and TEM specimens were successfully produced. Firstly, the Ca±P coatings were deposited on KCl crystal substrates from a HA target. The coated samples were then put into a beaker containing distilled water. As a result, the calcium phosphate coatings were peeled o after the dissolution of KCl crystal. The peeled o Ca±P coatings were examined under a TEM (JEOL TEM-100CX, Japan). FTIR was performed on a Nicolet FTIR 20SXB machine for characterizing various functional groups of the coatings, especially the hydroxyl and phosphate groups. The FTIR spectra were obtained using the transmittance mode from 4000 to 400 cm 1 . The surface morphology of coatings was examined by using a scanning electron microscope (SEM, Hitachi X-650, Hitachi, Japan). To prevent charging, the samples for SEM observations were coated with a thin layer of carbon. The elemental composition of coatings was determined by energy dispersive X-ray analysis (EDX). For the X-ray photoelectron spectroscopial (XPS) analysis, calcium phosphate coating/titanium samples were mounted on stainless steel stubs using double-sided adhesive and placed in a vacuum chamber. The vacuum chamber was then evacuated to minimum base pressure of 9 10 6 Pa. Using Mg Ka radiation (1253.6 eV), the coated surfaces were scanned over a range 0±1000 eV, with a pass energy of 100 eV. Relative atomic concentrations of all identi®ed elements were computed from multi-element data using peak areas and established elemental sensitivity factors. In addition, spectra for the elements were computer curve ®tted to yield the best binding-energy values. The binding energies for the elemental photoelectron peaks were corrected for charging using the adventitious carbon (C1s ) photoelectron peak at 284.6 eV as a reference. In order to determine the distribution of various elements in the thin Ca±P coatings, depth pro®ling was accomplished by sputter etching the coated samples for various periods of time and hence obtaining chemical composition of the coating at dierent depth. Using Ar ions, a differentially pumped ion gun (equivalent to 10 nm/ min sputtering rate) with a potential of 4 kV was
used to sputter etch the samples. Analysis of the sputtered surfaces was performed until the surface of the titanium substrate was reached. The adhesion of coatings to the substrate was evaluated by measuring the peel-o load (i.e. the detachment force) using a scanning-type scratch test system. 3. Results 3.1. Structural and morphology The XRD pattern of as-deposited monolayer calcium phosphate coatings from the HA target is shown in Fig. 1. As can be seen from this pattern, the coatings only exhibited a broad bump and no peaks other than those of titanium substrate were observed, indicating that as-deposited coatings were amorphous. The microstructure of as-deposited coatings was examined under TEM. The TEM analysis provided information on both the phases and the structure of coatings on a microscopic scale. Fig. 2(a) shows the bright ®eld TEM micrograph of the coating deposited with the ion beam energy of 1.2 keV and 40 mA. The corresponding selected area diraction (SAD) pattern is shown in Fig. 2(b). A single halo in the SAD pattern which con®rmed XRD results indicated that the as-deposited coatings were amorphous. FTIR was used to mainly characterize the hydroxyl and phosphate groups in the as-deposited coatings. According to Dasarathy et al. [16] and Walters et al. [17], in the molecular structure of HA, the phosphate group itself had a Td symmetry, resulting in four internal modes (symmetric stretch m1 : 956 cm 1 ; asymmetric stretch m2 : 430±460 cm 1 ; bending m3 : 1040±1090 cm 1 ; bending m4 : 575±610 cm 1 ; and the hydroxyl group with C1m had vibration modes at wavenumbers of 3570 and 630 cm 1 . Fig. 3 displays typical FTIR spectra of the Ar beam sputtering and N beam mixing Ca±P coating from HA target which were deposited on the KCl crystal substrate. In Fig. 3, the absorption bands at 1032 and 576 cm 1 were observed, indicating the existence of PO34 in the as-deposited Ca±P coating. No distinct absorption
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
367
Fig. 1. XRD pattern of as-deposited Ca±P coating from the HA target (Ar beam sputtering deposited monolayer coating, 60 mA cm 2 for 60 min).
Fig. 2. TEM micrograph of as-deposited Ca±P coating on the KCl crystal substrate (Ar beam sputtering deposited monolayer coating, 60 mA cm 2 for 60 min): (a) bright ®eld image; (b) corresponding SAD pattern.
bands of the hydroxyl group were observed. New absorption bands at 1494, 1425 and 939 cm 1 were present for CO23 , which was brought about by the deposition process. Similar results have also been found [12,13]. Fig. 4 shows the surface morphology of Ca±P coatings deposited on titanium substrate under
dierent conditions. The surface morphology of the coating was in¯uenced by the fabrication parameters. The coatings sputtering deposited by an Ar beam of 1200 eV and 40 mA cm 2 and the coatings produced by N mixing exhibited a clean and smooth surface morphology, while there were two dierent morphologies observed in
368
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
Fig. 3. FTIR spectrum of as-deposited Ca±P from the HA target (Ar beam sputtering deposited monolayer coating, 60 mA cm 60 min).
the coatings sputtering deposited by an Ar beam of 900 eV and 20 mA cm 2 . On the surface of this coating, some areas exhibited a clean and smooth surface morphology (zone A in Fig. 4(a)), while other areas exhibited a melt-like surface morphology (zone B in Fig. 4(b)). EDX spot analysis indicated that as-deposited coatings contained calcium and phosphorous. The semiquantative EDX analysis of relative amounts of calcium and phosphorous revealed that the Ca/P ratio of the coatings varied between 2.0 and 8.0. And the Ca/P ratio of the melt-like (zone B, Ca/ P 2.49) area was higher than that of the clean and smooth area (zone A, Ca/P 2.09) of the same coating. 3.2. Composition XPS results obtained from the surface of Ca±P coatings produced under dierent fabrication conditions are listed in Table 1. It can be seen that there was no dierence in the binding energy of the
2
for
elements in the coatings and the sputtering target, indicating that no valence variation occurred during the deposition process. With regard to the binding energy of carbon, the binding energy at 289 eV indicated the existence of CO23 . The surface Ca/P ratio of Ca±P coatings varied from 2 to 8, which was related to the fabrication parameters. Fig. 5 shows the relative atomic concentration of various elements in the sputtering deposited Ca±P coating as a function of sputter etching time. Based on these data, the coated sample could be divided into four distinct zones: the top surface, the thin Ca±P zone, the broad Ca±P±Ti intermixed zone, and the Ti substrate. In the top surface which was exposed to atmosphere, large quantities of carbon were found. Beneath the top surface, there was a thin Ca±P zone, in which no titanium element was detected. In the Ca±P±Ti intermixed zone, calcium phosphate and titanium coexisted. Under the Ca±P±Ti zone, only Ti was present.
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
369
Fig. 5. Compositional change in the sputtering deposited Ca±P coating from the XPS analysis (Ar beam sputtering deposited monolayer coating, 60 mA cm 2 for 60 min).
In addition, gradual compositional changes of the elements were achieved in the as-received coating. From the surface to the substrate, the relative amounts of calcium and phosphate elements gradually decreased, while the relative amounts of titanium element gradually increased. As for the coatings produced Ar sputtering and N mixing, after 30 min sputter etching, TiN was detected. This means that N has penetrated the full coating layer and reached the surface of the titanium substrate. And the combining of titanium atoms and nitrogen atoms led to the formation of TiN. The results obtained from the XPS analysis is listed in Table 2. In Table 2, binding energy at 455.9 eV is the binding energy of Ti peak for TiO, and 459 eV is the binding energy of Ti peak for TiO2 and TiN. And binding energy at 396.9 eV is the binding energy of N peak for TiN. In addition, there was no nitrogen element detected when the sputter etching time was shorter than 30 min. 3.3. Adhesion
Fig. 4. SEM micrographs of Ca±P coatings deposited on titanium substrate: (a) coating produced by an Ar of 900 eV and 20 mA cm 2 ; (b) coating produced by an Ar of 1200 eV and 40 mA cm 2 ; (c) coating produced by Ar sputtering and N mixing.
The adhesion of coatings to the substrate was evaluated using scratch tests. A scanning type scratch test system was used, where the stylus oscillated and moved forward while being pressed on the coating±substrate system with an increasing load. The radius of the stylus was 0.1 mm, the vibration amplitude was 2 mm, and the maximum applied load was 100 gf. In the current investigation, the peel-o load could not be obtained even at the maximum applied load, indicating that the coatings were very adherent to the substrate. The scratched samples were subsequently examined under the SEM. No delamination of the coating from the substrate was observed (Fig. 6). The high
370
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
Table 1 Binding energy and Ca/P ratio obtained from the coating surface by XPS analysesa Fabrication condition
Sputtering, 20 mA Sputtering, 40 mA Sputtering, 60 mA N mixing a
Binding energy (eV)
Ca/P ratio
C1s
Ca2p
P2p
284.6/289.0 284.6/288.5 284.6/288.9 284.6/288.7
347.4 347.4 347.2 347.4
133.1 133.1 132.9 133.2
2.45 1.59 3.7 7.87
Ca/P ratio for the target is 1.67.
Table 2 Binding energy obtained from the N mixing deposited coating surface by XPS analysesa Fabrication condition
Sputtering, 20 mA
Binding energy (eV) Ca2p
P2p
Ti2p
N1s
347.4
133.1
455.9/459
396.9
a The binding energy at 455.9 eV is the binding energy of Ti peak for TiO, binding energy at 459 eV is the binding energy of Ti peak for TiO2 and TiN, binding energy at 396.9 eV is the binding energy of N peak for TiN.
adhesion strength of the Ca±P coatings on titanium substrate may have resulted from the coating technique used. 4. Discussion XRD and TEM analyses showed that as-deposited Ca±P coatings were amorphous. This ``amorphous'' appearance is a direct result of the ion beam deposition technique. Because of the high vacuum and a relatively low temperature as compared to the plasma spraying coating method, there was not enough energy for the growth of nano-crystallites in the coatings during the deposition process, and hence the as-deposited coatings were shown to be amorphous. Using post-deposition heat treatment for the amorphous Ca±P coatings, the crystallinity of the Ca±P coatings could be ``increased'' [13]. This means that the heat treatment provides energy for the growth of nano-crystallites, and as a result, the crystallinity is shown to be ``increased''. In a separate study, part of the XRD peaks of HA was observed after heat treatment and some
crystal-like morphology was apparent in the surface of the coatings [18]. Variations for the phosphate group and the loss of the hydroxyl group in FTIR spectra were also caused by the ion beam deposition technique. During the deposition process, components of the target, such as hydroxide or oxygen and hydrogen, may not be transferred completely to the substrate (or anchored on the substrate surface and remained in the coatings) because of the requirement of maintaining a low pressure in this particular coating method. The variation in the Ca/P ratio among coatings was attributed to the fabrication parameter used. On the one hand, as phosphorus is very volatile, the low pressure during the sputtering process affected the anchoring of the sputtered phosphorus on the substrate. In addition, during the sputtering and the deposition process, one Ar ion sputtered dierent phosphorus and calcium atoms, meaning that the sputtering rate of calcium atoms and that of phosphorus atoms were dierent. On the other hand, according to Cotell [19], a possible cause was the substitution of carbonate group for the phosphate groups in the HA molecule. The process
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
371
5. Conclusions
Fig. 6. SEM micrograph of the scratch in the sputtering deposited Ca±P coating (Ar beam sputtering deposited monolayer coating, 60 mA cm 2 for 60 min).
of ion beam sputtering HA target had been studied using theoretical analysis [20]. The results showed that, because of the collision of sputtering ions and HA molecules, defects such as lattice displacement and vacancy were produced in the coating. These defects made it possible for the substitution of carbonate group for the phosphate groups in the HA molecule. As mentioned earlier, one of the advantages of using the ion beam sputtering deposition technique is the production of thin coatings with high density and excellent adhesion. The scratch tests showed that the Ca±P coating adhered well to the titanium substrate. According to results obtained from XPS depth pro®ling, the Ca±P/titanium system can be divided into four distinct zones: the top surface, the Ca±P zone, the Ca±P±Ti intermixed zone, and the Ti substrate. As for the composition of coatings, gradual compositional changes were achieved. During the sputtering deposition processes, a thick Ca±P±Ti intermixed zone was formed due to the diusion of coating and the substrate atoms. In this zone, on the one hand, Ti atoms from the substrate were diusing to the Ca± P coating; on the other hand, atoms of the Ca±P coating were also diusing to the Ti substrate. This intermixed zone is believed to be bene®cial in forming a compositional gradient. Such an intermixed zone can provide good adhesion of the deposited coatings to the substrate.
Dense and homogeneous Ca±P coatings were successfully produced on titanium substrate using the ion beam deposition technique. The results showed that the as-deposited coatings were amorphous. In comparison with the HA ceramic target, some variations in chemical composition of the coatings were brought about in the deposition process, such as the distortion of the phosphate lattice, loss of hydroxyl groups, and the incorporation of CO23 . However, no valence variation of the elements in the coatings was found. Depth pro®ling of the coatings revealed four distinct zones: the top surface, the thin coating zone, the intermixed zone of coating and substrate, and the substrate. Scratch tests showed that the coatings adhered well to the substrate. The existence of the broad intermixed zone may have contributed to the good adhesion of coatings to the substrate. TiN was found on the surface of titanium substrate of the N mixing deposited coatings, this proved that the choice of the energy and mixing dosage of N in this investigation was reasonable. The subsequent in¯uence of the formation of TiN on the Ca±P coatings will be studied by using other analysis methods. Acknowledgements The work reported in this paper was supported by the National 863 Program of New Technology and the National Natural Science Foundation of China. References [1] L. Feendral, K. de Groot, in: K. de Groot (Ed.), Bioceramic of Calcium Phosphate, CRC Press, Boca Raton, FL, USA, 1982. [2] G. Daculsi, R.Z. LeGeros, C. Deudon, Scan. Microsc. 4 (1990) 309. [3] R. Bell, O. Beirne, J. Oral Maxillofac Surg. 46 (1988) 589. [4] J.F. Kay, J. Oral Implants 14 (1988) 43. [5] W.R. Lace®eld, in: P. Ducheyne, J.E. Lemons (Eds.), Bioceramics: Material Characteristics Versus In Vivo Behavior, New York, Ann. Acad. Sci. (1988) 72. [6] H. Monma, J. Ceram. Soc. Jpn., Int. ed. 101 (1993) 718.
372
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
[7] J.G.C. Wolke, V.J.P.C. der Waerden, K. de Groot, J.A. Jansen, Biomaterials 18 (1997) 483. [8] R.K. Singh, F. Qian, V. Nagabushnam, R.M. Damodaran, B. Moudgil, Biomaterials 15 (1994) 522. [9] C.K. Wang, J.H. Chern Lin, C.P. Ju, H.C. Ong, R.P.H. Chang, Biomaterials 18 (1997) 1331. [10] H.B. Wen, J.R. De Wijin, F.Z. Cui, K. De Groot, Biomaterials 19 (1998) 215. [11] Q.Y. Zhang, Z.H. Long, C.S. Ren, B.H. Guo, D.Q. Xu, T.C. Ma, Surf. Coat. Technol. 103±104 (1998) 195. [12] J.L. Ong, L.C. Lucas, W.R. Lace®eld, E.D. Rigney, Biomaterials 13 (1992) 249. [13] J.L. Ong, L.C. Lucas, Biomaterials 15 (1994) 337. [14] Q. Zhao, G.J. Zhai, D.H. Ng, X.Z. Zhang, Z.Q. Chen, Biomaterials 20 (1999) 595.
[15] C.X. Wang, Z.Q. Chen, L.M. Guan, Z.Y. Liu, P.L. Wang, S.X. Zheng, X.D. Liao, Surf. Coat. Technol. 130 (1) (2000) 39. [16] H. Dasarathy, C. Riley, H.D. Coble, J. Biomed. Mater. Res. 27 (1993) 477. [17] M.A. Walters, Y.C. Leung, N.C. Blumenthal, R.Z. Legeros, K.A. Konsker, J. Inorg. Biochem. 39 (1990) 193. [18] Z.Q. Chen, C.X. Wang, L.M. Guan, M. Wang, in: Proceedings of PFAM-8, Singapore, 1999. [19] M.C. Cotell, Appl. Surf. Sci. 69 (1993) 140. [20] C.X. Wang, Z.Q. Chen, H.L. Zhang, in: Proceedings of the 20th Annual International Conference of the IEEE, Hong Kong, 1998, p. 2890.
www.elsevier.com/locate/nimb
Fabrication and characterization of graded calcium phosphate coatings produced by ion beam sputtering/mixing deposition C.X. Wang a
c
a,b,*
, Z.Q. Chen a, L.M. Guan a, M. Wang b, Z.Y. Liu c, P.L. Wang c
Department of Dental Materials, College of Stomatology, West China University of Medical Sciences, Chengdu 610041, Sichuan, China b School of Mechanical and Production Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore Laboratory of Radiation Physics and Technology, Institute of Nuclear Science and Technology, Sichuan University, Chengdu 610064, Sichuan, China Received 2 January 2001; received in revised form 23 February 2001
Abstract Ion beam sputtering/mixing deposition was used to produce thin calcium phosphate coatings on titanium substrate from the hydroxyapatite target. It was found that as-deposited coatings were amorphous. No distinct absorption band of the hydroxyl group was observed in FTIR spectra of the coatings but new absorption bands were present for CO23 , which was brought about during the deposition process. Scanning electron microscopy revealed that the deposited coatings had a uniform and dense structure. The calcium to phosphorous ratio of these coatings varied between 2.0 and 8.0. Analyses of XPS data revealed that the coating could be divided into four distinctive zones, and a graded structure was achieved in the as-received coating. Scratch tests showed that the coatings adhered well to the substrate. Ó 2001 Elsevier Science B.V. All rights reserved. Keywords: Ion beam sputtering deposition; Hydroxyapatite; Calcium phosphate coating; Titanium substrate; Characterization
1. Introduction Hydroxyapatite (HA) is a biocompatible and bioactive material. It is well known for its osteoconductivity [1,2]. However, the brittle nature of
*
Corresponding author. Tel.: +65-790-4963; fax: +65-7911859. E-mail address: [email protected] (C.X. Wang).
HA ceramic limits its clinical use in load-bearing situations such as total hip replacement. HA, nevertheless, has been used as a coating for metal implants, because it can speed up bone healing adjacent to implants and form a strong chemical bond with bone [3,4]. Currently, commercially available coated metal implants are manufactured by using the plasma spray technique for depositing HA coatings. However, problems such as low bond strength
0168-583X/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 5 8 3 X ( 0 1 ) 0 0 5 8 5 - 7
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
between the coating and the substrate and nonuniformity across the thickness of the coating are often encountered with plasma sprayed coatings [5]. A variety of other coating techniques, such as electrochemical deposition [6], radio-frequency magnetron sputtering [7], excimer laser deposition [8], pulsed laser deposition [9] and dipping [10], have also been investigated for producing HA coatings on metallic substrates. Since the mid1970s, many surface modi®cation techniques that are based on ion implantation, such as ion beam assisted deposition (IBAD), ion beam deposition (IBD), ion beam mixing (IBM), and techniques that are based on plasma-assisted ion implantation such as plasma source ion implantation (PSII) and plasma immersion ion implantation (PIII), have been developed rapidly and are now widely used to modify the surface of materials including metals, ceramics and polymers [11]. Ion beam sputter deposition was investigated as a method for producing biocompatible ceramic coatings on metallic implants due to its advantages which include the production of thin coatings with high density and superior adhesion [12,13]. In this process, the ionized-argon gas was used to sputter atoms from a ceramic target. The sputtered atoms built up on the metallic substrate that was placed in the path of the sputtered material. Both argon ion beam and nitrogen ion beam can be used for the ion beam bombardment. It was found that grafting ±NH 2 amidogen radicals on Al2 O3 ceramic using ion implantation technique could improve the biocompatibility of this ceramic [14]. So if using nitrogen ion beam to bombard the Ca±P coatings, it is possible to further improve the bone bonding between the coating and bone to expedite the wound healing. Moreover, during the bombardment process, nitrogen ions could penetrate into the whole coating layer and bond with the titanium substrate to form Tix Ny , which is bene®cial to improve the properties of coatings. In this investigation, the ion beam sputtering/mixing deposition technique was used to produce thin calcium phosphate coatings on titanium substrate. The emphasis was placed on the structural and chemical characterizations of the coatings produced by nitrogen ion beam mixing in this paper.
365
2. Materials and methods A commercially available pure titanium substrate (plates of dimensions 20 mm 20 mm 1 mm) was mechanically polished and ultrasonically cleaned with acetone and alcohol. HA powder was synthesized using the wet method. HA powder disks, which were to be used as the ion beam sputtering target, were cold-pressed and sintered at 1150°C for 3 h. X-ray diraction patterns of HA disks matched that of the standard synthetic hydroxyapatite (PDFSM #9-0432), indicating that the disks had well crystallized HA with a microstructure of purely random grain orientation. The ion beam sputtering deposition system has been described previously [15]. It mainly consisted of one Kaufman ion source and one Freeman ion source, one target holder and one rotatable sample holder in the path of both ion beams. The deposition chamber was evacuated to a base pressure of 2:8±3:7 10 4 Pa. Prior to deposition, etching substrates with 800 eV and 40 mA cm 2 argon ions for 30 min was performed to clean the surface of titanium substrate. The energetic ion beam was produced by ionizing high purity argon gas (99.999% pure). After cleaning, the stage was rotated so that the substrates were placed in the path of the sputtered atoms. The sputtering deposited monolayer coatings were produced using an Ar beam of 900 eV and 20 mA cm 2 for 3 h, 1200 eV and 40 mA cm 2 for 90 min and 1500 eV and 60 mA cm 2 for 60 min, respectively. For the Ar beam sputtering and N beam mixing deposited Ca±P coatings, after sputtering the HA target with 1200 eV and 40 mA cm 2 Ar beam for 1.5 h, the N beam with 60 KeV was used to homogenize the coating and its dosage was 2:0 1016 ions cm 2 . The thickness of all the coatings is about 300 nm. X-ray diraction (XRD) was employed to analyze the structure of as-sputtered coatings. A Rigaiku D/max-sA X-ray diractometer with Cu Ka radiation at 40 KeV and 50±80 mA was used. Transmission electron microscopy (TEM) was also used to examine the microstructure of as-deposited coatings. Because of the brittle nature of thin Ca±P coatings, it was very dicult to prepare suitable specimens for TEM observation directly from coatings with titanium substrate by using an ion
366
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
beam thinning apparatus. An alternative method was used and TEM specimens were successfully produced. Firstly, the Ca±P coatings were deposited on KCl crystal substrates from a HA target. The coated samples were then put into a beaker containing distilled water. As a result, the calcium phosphate coatings were peeled o after the dissolution of KCl crystal. The peeled o Ca±P coatings were examined under a TEM (JEOL TEM-100CX, Japan). FTIR was performed on a Nicolet FTIR 20SXB machine for characterizing various functional groups of the coatings, especially the hydroxyl and phosphate groups. The FTIR spectra were obtained using the transmittance mode from 4000 to 400 cm 1 . The surface morphology of coatings was examined by using a scanning electron microscope (SEM, Hitachi X-650, Hitachi, Japan). To prevent charging, the samples for SEM observations were coated with a thin layer of carbon. The elemental composition of coatings was determined by energy dispersive X-ray analysis (EDX). For the X-ray photoelectron spectroscopial (XPS) analysis, calcium phosphate coating/titanium samples were mounted on stainless steel stubs using double-sided adhesive and placed in a vacuum chamber. The vacuum chamber was then evacuated to minimum base pressure of 9 10 6 Pa. Using Mg Ka radiation (1253.6 eV), the coated surfaces were scanned over a range 0±1000 eV, with a pass energy of 100 eV. Relative atomic concentrations of all identi®ed elements were computed from multi-element data using peak areas and established elemental sensitivity factors. In addition, spectra for the elements were computer curve ®tted to yield the best binding-energy values. The binding energies for the elemental photoelectron peaks were corrected for charging using the adventitious carbon (C1s ) photoelectron peak at 284.6 eV as a reference. In order to determine the distribution of various elements in the thin Ca±P coatings, depth pro®ling was accomplished by sputter etching the coated samples for various periods of time and hence obtaining chemical composition of the coating at dierent depth. Using Ar ions, a differentially pumped ion gun (equivalent to 10 nm/ min sputtering rate) with a potential of 4 kV was
used to sputter etch the samples. Analysis of the sputtered surfaces was performed until the surface of the titanium substrate was reached. The adhesion of coatings to the substrate was evaluated by measuring the peel-o load (i.e. the detachment force) using a scanning-type scratch test system. 3. Results 3.1. Structural and morphology The XRD pattern of as-deposited monolayer calcium phosphate coatings from the HA target is shown in Fig. 1. As can be seen from this pattern, the coatings only exhibited a broad bump and no peaks other than those of titanium substrate were observed, indicating that as-deposited coatings were amorphous. The microstructure of as-deposited coatings was examined under TEM. The TEM analysis provided information on both the phases and the structure of coatings on a microscopic scale. Fig. 2(a) shows the bright ®eld TEM micrograph of the coating deposited with the ion beam energy of 1.2 keV and 40 mA. The corresponding selected area diraction (SAD) pattern is shown in Fig. 2(b). A single halo in the SAD pattern which con®rmed XRD results indicated that the as-deposited coatings were amorphous. FTIR was used to mainly characterize the hydroxyl and phosphate groups in the as-deposited coatings. According to Dasarathy et al. [16] and Walters et al. [17], in the molecular structure of HA, the phosphate group itself had a Td symmetry, resulting in four internal modes (symmetric stretch m1 : 956 cm 1 ; asymmetric stretch m2 : 430±460 cm 1 ; bending m3 : 1040±1090 cm 1 ; bending m4 : 575±610 cm 1 ; and the hydroxyl group with C1m had vibration modes at wavenumbers of 3570 and 630 cm 1 . Fig. 3 displays typical FTIR spectra of the Ar beam sputtering and N beam mixing Ca±P coating from HA target which were deposited on the KCl crystal substrate. In Fig. 3, the absorption bands at 1032 and 576 cm 1 were observed, indicating the existence of PO34 in the as-deposited Ca±P coating. No distinct absorption
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
367
Fig. 1. XRD pattern of as-deposited Ca±P coating from the HA target (Ar beam sputtering deposited monolayer coating, 60 mA cm 2 for 60 min).
Fig. 2. TEM micrograph of as-deposited Ca±P coating on the KCl crystal substrate (Ar beam sputtering deposited monolayer coating, 60 mA cm 2 for 60 min): (a) bright ®eld image; (b) corresponding SAD pattern.
bands of the hydroxyl group were observed. New absorption bands at 1494, 1425 and 939 cm 1 were present for CO23 , which was brought about by the deposition process. Similar results have also been found [12,13]. Fig. 4 shows the surface morphology of Ca±P coatings deposited on titanium substrate under
dierent conditions. The surface morphology of the coating was in¯uenced by the fabrication parameters. The coatings sputtering deposited by an Ar beam of 1200 eV and 40 mA cm 2 and the coatings produced by N mixing exhibited a clean and smooth surface morphology, while there were two dierent morphologies observed in
368
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
Fig. 3. FTIR spectrum of as-deposited Ca±P from the HA target (Ar beam sputtering deposited monolayer coating, 60 mA cm 60 min).
the coatings sputtering deposited by an Ar beam of 900 eV and 20 mA cm 2 . On the surface of this coating, some areas exhibited a clean and smooth surface morphology (zone A in Fig. 4(a)), while other areas exhibited a melt-like surface morphology (zone B in Fig. 4(b)). EDX spot analysis indicated that as-deposited coatings contained calcium and phosphorous. The semiquantative EDX analysis of relative amounts of calcium and phosphorous revealed that the Ca/P ratio of the coatings varied between 2.0 and 8.0. And the Ca/P ratio of the melt-like (zone B, Ca/ P 2.49) area was higher than that of the clean and smooth area (zone A, Ca/P 2.09) of the same coating. 3.2. Composition XPS results obtained from the surface of Ca±P coatings produced under dierent fabrication conditions are listed in Table 1. It can be seen that there was no dierence in the binding energy of the
2
for
elements in the coatings and the sputtering target, indicating that no valence variation occurred during the deposition process. With regard to the binding energy of carbon, the binding energy at 289 eV indicated the existence of CO23 . The surface Ca/P ratio of Ca±P coatings varied from 2 to 8, which was related to the fabrication parameters. Fig. 5 shows the relative atomic concentration of various elements in the sputtering deposited Ca±P coating as a function of sputter etching time. Based on these data, the coated sample could be divided into four distinct zones: the top surface, the thin Ca±P zone, the broad Ca±P±Ti intermixed zone, and the Ti substrate. In the top surface which was exposed to atmosphere, large quantities of carbon were found. Beneath the top surface, there was a thin Ca±P zone, in which no titanium element was detected. In the Ca±P±Ti intermixed zone, calcium phosphate and titanium coexisted. Under the Ca±P±Ti zone, only Ti was present.
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
369
Fig. 5. Compositional change in the sputtering deposited Ca±P coating from the XPS analysis (Ar beam sputtering deposited monolayer coating, 60 mA cm 2 for 60 min).
In addition, gradual compositional changes of the elements were achieved in the as-received coating. From the surface to the substrate, the relative amounts of calcium and phosphate elements gradually decreased, while the relative amounts of titanium element gradually increased. As for the coatings produced Ar sputtering and N mixing, after 30 min sputter etching, TiN was detected. This means that N has penetrated the full coating layer and reached the surface of the titanium substrate. And the combining of titanium atoms and nitrogen atoms led to the formation of TiN. The results obtained from the XPS analysis is listed in Table 2. In Table 2, binding energy at 455.9 eV is the binding energy of Ti peak for TiO, and 459 eV is the binding energy of Ti peak for TiO2 and TiN. And binding energy at 396.9 eV is the binding energy of N peak for TiN. In addition, there was no nitrogen element detected when the sputter etching time was shorter than 30 min. 3.3. Adhesion
Fig. 4. SEM micrographs of Ca±P coatings deposited on titanium substrate: (a) coating produced by an Ar of 900 eV and 20 mA cm 2 ; (b) coating produced by an Ar of 1200 eV and 40 mA cm 2 ; (c) coating produced by Ar sputtering and N mixing.
The adhesion of coatings to the substrate was evaluated using scratch tests. A scanning type scratch test system was used, where the stylus oscillated and moved forward while being pressed on the coating±substrate system with an increasing load. The radius of the stylus was 0.1 mm, the vibration amplitude was 2 mm, and the maximum applied load was 100 gf. In the current investigation, the peel-o load could not be obtained even at the maximum applied load, indicating that the coatings were very adherent to the substrate. The scratched samples were subsequently examined under the SEM. No delamination of the coating from the substrate was observed (Fig. 6). The high
370
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
Table 1 Binding energy and Ca/P ratio obtained from the coating surface by XPS analysesa Fabrication condition
Sputtering, 20 mA Sputtering, 40 mA Sputtering, 60 mA N mixing a
Binding energy (eV)
Ca/P ratio
C1s
Ca2p
P2p
284.6/289.0 284.6/288.5 284.6/288.9 284.6/288.7
347.4 347.4 347.2 347.4
133.1 133.1 132.9 133.2
2.45 1.59 3.7 7.87
Ca/P ratio for the target is 1.67.
Table 2 Binding energy obtained from the N mixing deposited coating surface by XPS analysesa Fabrication condition
Sputtering, 20 mA
Binding energy (eV) Ca2p
P2p
Ti2p
N1s
347.4
133.1
455.9/459
396.9
a The binding energy at 455.9 eV is the binding energy of Ti peak for TiO, binding energy at 459 eV is the binding energy of Ti peak for TiO2 and TiN, binding energy at 396.9 eV is the binding energy of N peak for TiN.
adhesion strength of the Ca±P coatings on titanium substrate may have resulted from the coating technique used. 4. Discussion XRD and TEM analyses showed that as-deposited Ca±P coatings were amorphous. This ``amorphous'' appearance is a direct result of the ion beam deposition technique. Because of the high vacuum and a relatively low temperature as compared to the plasma spraying coating method, there was not enough energy for the growth of nano-crystallites in the coatings during the deposition process, and hence the as-deposited coatings were shown to be amorphous. Using post-deposition heat treatment for the amorphous Ca±P coatings, the crystallinity of the Ca±P coatings could be ``increased'' [13]. This means that the heat treatment provides energy for the growth of nano-crystallites, and as a result, the crystallinity is shown to be ``increased''. In a separate study, part of the XRD peaks of HA was observed after heat treatment and some
crystal-like morphology was apparent in the surface of the coatings [18]. Variations for the phosphate group and the loss of the hydroxyl group in FTIR spectra were also caused by the ion beam deposition technique. During the deposition process, components of the target, such as hydroxide or oxygen and hydrogen, may not be transferred completely to the substrate (or anchored on the substrate surface and remained in the coatings) because of the requirement of maintaining a low pressure in this particular coating method. The variation in the Ca/P ratio among coatings was attributed to the fabrication parameter used. On the one hand, as phosphorus is very volatile, the low pressure during the sputtering process affected the anchoring of the sputtered phosphorus on the substrate. In addition, during the sputtering and the deposition process, one Ar ion sputtered dierent phosphorus and calcium atoms, meaning that the sputtering rate of calcium atoms and that of phosphorus atoms were dierent. On the other hand, according to Cotell [19], a possible cause was the substitution of carbonate group for the phosphate groups in the HA molecule. The process
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
371
5. Conclusions
Fig. 6. SEM micrograph of the scratch in the sputtering deposited Ca±P coating (Ar beam sputtering deposited monolayer coating, 60 mA cm 2 for 60 min).
of ion beam sputtering HA target had been studied using theoretical analysis [20]. The results showed that, because of the collision of sputtering ions and HA molecules, defects such as lattice displacement and vacancy were produced in the coating. These defects made it possible for the substitution of carbonate group for the phosphate groups in the HA molecule. As mentioned earlier, one of the advantages of using the ion beam sputtering deposition technique is the production of thin coatings with high density and excellent adhesion. The scratch tests showed that the Ca±P coating adhered well to the titanium substrate. According to results obtained from XPS depth pro®ling, the Ca±P/titanium system can be divided into four distinct zones: the top surface, the Ca±P zone, the Ca±P±Ti intermixed zone, and the Ti substrate. As for the composition of coatings, gradual compositional changes were achieved. During the sputtering deposition processes, a thick Ca±P±Ti intermixed zone was formed due to the diusion of coating and the substrate atoms. In this zone, on the one hand, Ti atoms from the substrate were diusing to the Ca± P coating; on the other hand, atoms of the Ca±P coating were also diusing to the Ti substrate. This intermixed zone is believed to be bene®cial in forming a compositional gradient. Such an intermixed zone can provide good adhesion of the deposited coatings to the substrate.
Dense and homogeneous Ca±P coatings were successfully produced on titanium substrate using the ion beam deposition technique. The results showed that the as-deposited coatings were amorphous. In comparison with the HA ceramic target, some variations in chemical composition of the coatings were brought about in the deposition process, such as the distortion of the phosphate lattice, loss of hydroxyl groups, and the incorporation of CO23 . However, no valence variation of the elements in the coatings was found. Depth pro®ling of the coatings revealed four distinct zones: the top surface, the thin coating zone, the intermixed zone of coating and substrate, and the substrate. Scratch tests showed that the coatings adhered well to the substrate. The existence of the broad intermixed zone may have contributed to the good adhesion of coatings to the substrate. TiN was found on the surface of titanium substrate of the N mixing deposited coatings, this proved that the choice of the energy and mixing dosage of N in this investigation was reasonable. The subsequent in¯uence of the formation of TiN on the Ca±P coatings will be studied by using other analysis methods. Acknowledgements The work reported in this paper was supported by the National 863 Program of New Technology and the National Natural Science Foundation of China. References [1] L. Feendral, K. de Groot, in: K. de Groot (Ed.), Bioceramic of Calcium Phosphate, CRC Press, Boca Raton, FL, USA, 1982. [2] G. Daculsi, R.Z. LeGeros, C. Deudon, Scan. Microsc. 4 (1990) 309. [3] R. Bell, O. Beirne, J. Oral Maxillofac Surg. 46 (1988) 589. [4] J.F. Kay, J. Oral Implants 14 (1988) 43. [5] W.R. Lace®eld, in: P. Ducheyne, J.E. Lemons (Eds.), Bioceramics: Material Characteristics Versus In Vivo Behavior, New York, Ann. Acad. Sci. (1988) 72. [6] H. Monma, J. Ceram. Soc. Jpn., Int. ed. 101 (1993) 718.
372
C.X. Wang et al. / Nucl. Instr. and Meth. in Phys. Res. B 179 (2001) 364±372
[7] J.G.C. Wolke, V.J.P.C. der Waerden, K. de Groot, J.A. Jansen, Biomaterials 18 (1997) 483. [8] R.K. Singh, F. Qian, V. Nagabushnam, R.M. Damodaran, B. Moudgil, Biomaterials 15 (1994) 522. [9] C.K. Wang, J.H. Chern Lin, C.P. Ju, H.C. Ong, R.P.H. Chang, Biomaterials 18 (1997) 1331. [10] H.B. Wen, J.R. De Wijin, F.Z. Cui, K. De Groot, Biomaterials 19 (1998) 215. [11] Q.Y. Zhang, Z.H. Long, C.S. Ren, B.H. Guo, D.Q. Xu, T.C. Ma, Surf. Coat. Technol. 103±104 (1998) 195. [12] J.L. Ong, L.C. Lucas, W.R. Lace®eld, E.D. Rigney, Biomaterials 13 (1992) 249. [13] J.L. Ong, L.C. Lucas, Biomaterials 15 (1994) 337. [14] Q. Zhao, G.J. Zhai, D.H. Ng, X.Z. Zhang, Z.Q. Chen, Biomaterials 20 (1999) 595.
[15] C.X. Wang, Z.Q. Chen, L.M. Guan, Z.Y. Liu, P.L. Wang, S.X. Zheng, X.D. Liao, Surf. Coat. Technol. 130 (1) (2000) 39. [16] H. Dasarathy, C. Riley, H.D. Coble, J. Biomed. Mater. Res. 27 (1993) 477. [17] M.A. Walters, Y.C. Leung, N.C. Blumenthal, R.Z. Legeros, K.A. Konsker, J. Inorg. Biochem. 39 (1990) 193. [18] Z.Q. Chen, C.X. Wang, L.M. Guan, M. Wang, in: Proceedings of PFAM-8, Singapore, 1999. [19] M.C. Cotell, Appl. Surf. Sci. 69 (1993) 140. [20] C.X. Wang, Z.Q. Chen, H.L. Zhang, in: Proceedings of the 20th Annual International Conference of the IEEE, Hong Kong, 1998, p. 2890.