Appl. Phys. A 76, 251–256 (2003)

Applied Physics A

DOI: 10.1007/s003390201423

Materials Science & Processing

✉ ´ j.m. fernandez-pradas g. sardin j.l. morenza

Inhomogeneity of calcium phosphate coatings deposited by laser ablation at high deposition rate Universitat de Barcelona, Departament de F´ısica Aplicada i Òptica, Av. Diagonal 647, 08028 Barcelona, Spain

Received: 22 November 2001/Accepted: 12 March 2002 Published online: 5 July 2002 • © Springer-Verlag 2002 ABSTRACT Calcium phosphate coatings were deposited with a KrF excimer laser onto titanium alloy to study their homogeneity. Deposition was performed at a high deposition rate under a water vapour atmosphere of 45 Pa and at a substrate temperature of 575 ◦ C. Samples were also submitted to annealing under the same conditions of deposition for different times just after deposition. The effects of the annealing were also investigated. The morphology of the coatings was studied by scanning electron microscopy. Their structure and phase distribution was analysed by X-ray diffractometry and infrared and micro-Raman spectroscopies. Besides the non-uniform thickness, the results reveal an inhomogeneity in the spatial distribution of calcium phosphate phases in the coatings. The phase distribution can be almost completely correlated with the deposition rate. High deposition rates (0.5 nm/pulse) occurring in the centre of deposition results in the formation of amorphous calcium phosphate, while lower deposition rates favour the presence of hydroxyapatite and alpha tricalcium phosphate. At intermediate deposition rates, beta tricalcium phosphate is found, probably because the superimposed effect of energetic particles bombardment. The annealing process promotes the crystallisation of the amorphous material. The importance of the deposition rate in the phases obtained is stated after comparing these results with a previous work where homogeneous hydroxyapatite coatings were obtained under the same conditions of laser fluence, temperature and pressure, but at lower deposition rates. PACS 81.15.Fg;

1

68.55.-a; 87.68.+z

Introduction

Hydroxyapatite (HA = Ca10 (PO4 )6 (OH)2 ) is a calcium phosphate very similar to the biologic mineral phase of bone [1]. For that reason it was proposed as a biomaterial for osseous repair and reconstruction. HA indeed presents bioactive properties like most of the calcium phosphates. New bone bonds directly to HA implants [2]. Moreover, HA is the least resorbable calcium phosphate material [3], therefore it confers stability to the implant [4]. However, the brittle nature of ✉ Fax: +34-93/402-1138, E-mail: [email protected]

the HA ceramic does not permit its application in load-bearing sites [5]. In these cases, metallic implants can be coated with HA so as to take advantage of its bioactivity preserving the mechanical performance of the metallic structure [5]. Among different techniques for HA deposition, pulsed laser deposition (PLD) has been studied in the last decade with the intention of producing thin, dense and well-adhered HA crystalline coatings [6–15]. Different type of lasers have been used, and research has been mainly centred in the influence on the coatings structure of the deposition parameters such as the nature of the reactive atmosphere, the pressure of this atmosphere and the temperature of deposition. The effects of other parameters like the target-substrate distance or the laser fluence have also been investigated. The results of these studies show that it is possible to obtain calcium phosphate coatings with specific composition and structure by controlling the parameters of deposition. In particular, dense coatings of highly crystalline HA can be obtained. In vitro studies of these coatings have demonstrated that they are stable even in osteoclastic cell cultures [16, 17]. Moreover, they promote the formation of a mineralised apatite layer well adhered on its surface in simulated body fluid solution and in osteoblast cell cultures [17, 18]. Laser ablation produces the emission of neutral and ionised species which constitute a luminous plasma called plume [19, 20]. The energy and amount of the travelling species have a narrow angular distribution that can lead to inhomogeneities in the coatings, the most evident concerning coating thickness [21]. This effect was already observed in the first work on HA deposition [6], but it has not received more attention beyond that. In fact, the non-uniformity of the coating thickness is only one of the consequences of the nonuniformity of the deposition rate, because this can also affect the structure and composition homogeneity of the coatings. In particular, it was shown that high deposition rates favour the presence of other calcium phosphates in front of HA in the coatings [22]. The work presented here is a study of the spatial distribution of the calcium phosphate phases in a PLD coating. A high deposition rate has been selected for coating deposition in order to obtain significant differences of deposition rate, and thus in composition. The evolution of the calcium phosphate phases in the coating with a post-annealing process in situ has also been studied. To clarify the role of the deposition rate on the composition of the coatings, the re-

252

Applied Physics A – Materials Science & Processing

sults are compared with those of coatings obtained at lower deposition rates [23]. 2

Experimental

HA targets were produced by compressing commercially available HA powder at a pressure of 240 MPa. The resulting target density was 1.5 g/cm3 . Two 1 × 1 cm2 foils of Ti-6Al-4V were coated in each deposition process. Prior to deposition, the substrates were polished with SiC paper up to a final roughness of 0.03 µm and cleaned in a sequence of ultrasonic baths of trichloroethylene, acetone and ethanol. Coatings were deposited in a vacuum chamber. A pulsed KrF excimer laser with a wavelength of 248 nm was used for ablation. The laser gave pulses of 34 ns and 130 mJ at a repetition rate of 10 Hz. The laser beam was focused on the rotating HA target at an angle of 45 ◦ C to get a fluence of 2.2 J/cm2 . This fluence was selected because fluences around 2 J/cm2 avoid evolutions of the deposition rate during deposition [24]. The two Ti-6Al-4V foils were fixed on a copper substrateheater placed at 4 cm in front of the target. The first foil was located just in front of the laser spot on the target so the ablation plume was centred on it. The second foil was placed next to the first one (Fig. 1a). The substrate temperature was measured on a substrate replica placed on the same substrate-heater by using a K-type thermocouple attached to its surface. Deposition was performed at a substrate temperature of 575 ◦ C under a water vapour atmosphere of 45 Pa. Coatings were grown as the result of 4000 laser pulses. Just after deposition, some samples were annealed in situ under the same conditions of temperature and water atmosphere for 8, 23 or 38 min. Samples were cooled down to room temperature under the water atmosphere by simply switching off the substrate-heater source.

The morphology of the coatings was observed by scanning electron microscopy (SEM). X-ray diffractometry (XRD), together with infrared spectroscopy (IR), were used to study the phase composition of the coatings in a macroscopic scale. These data were complemented with micro-Raman spectroscopy that allowed us to analyse the spatial distribution of the calcium phosphate phases in the coatings with a lateral resolution of around 1 µm. 3

Results

The coatings have a maximum thickness of 2.0 µm at the centre of the plume in foil 1. At this place the deposition rate was 0.5 nm/pulse. The coating thickness decays as a cos2 θ function in foil 1 down to 1 µm (Fig. 1b). The coatings in foil 2 are more homogeneous and their thickness varies between 1 µm and 0.8 µm. Thus the minimum deposition rate collected on the substrate is 0.2 nm/pulse. When the coatings were examined with polarised light in the optical microscope, an elliptical ring with bright contrast of about 1 mm wide was detected in foil 1. This ring is identified as zone II, and its inner and outer parts in foil 1 as zone I and zone III respectively. The distribution of these zones is depicted in Fig. 1a. Zone II presents bright regions at the microscope that seem to correspond to oriented crystalline domains (see Fig. 1a), while the rest of the coating is seen in dark. The surface morphology of the as-deposited coating shows in the entire surface of both foils a similar granular structure with some droplets (Fig. 2). This morphology is typical for coatings deposited with a KrF excimer laser [23]. The annealing process does not significantly alter the morphology. The X-ray diffractograms of the coatings show different characteristics depending whether they correspond to foil types 1 or 2 (Fig. 3). The diffractogram of the as-deposited coating on foil 1 presents small and broad peaks that are hardly distinguishable from the background. These features indicate that the deposited material has low crystalline qualities. However, some peaks can be intuitively glimpsed looking to the evolution of the diffractograms with the annealing time. Thus, the presence of some tricalcium phosphate (TCP = Ca3 (PO4 )2 ), in both alpha (α-TCP) and beta (β-TCP) phases, and HA can be detected. As seen, the peaks of these phases are higher and better defined as the time of annealing

a

b a Arrangement of the substrate constituted by two foils. The centre of the ablation plume falls in foil 1. Three zones are distinguished in foil 1. The inserted photograph shows zone II observed at the optical microscope with polarised light. b Thickness profile of the coating measured along the dashed line indicated in a FIGURE 1

FIGURE 2 SEM image of the surface of the as-deposited coating. The sample is tilted 45 ◦ C

´ FERN ANDEZ - PRADAS et al.

Inhomogeneity of calcium phosphate coatings deposited by laser ablation

253

(d)

(c)

(b) (d)

(a)

(c) (b) (a)

(d) (d)

(c)

(c) (b)

(b)

(a)

(a)

FIGURE 3 X-ray diffractograms of (a) the as-deposited coatings, and the annealed coatings for (b) 8, (c) 23, and (d) 38 min, corresponding to the two foils

increases. The as-deposited coating on foil 2 presents reflections of HA and α-TCP. These reflections are also increased in the case of annealed coatings. This indicates that the asdeposited material is not fully crystalline, and that the annealing process enhances the crystallisation of the coatings. The IR reflectance spectra of the samples are shown in Fig. 4. The spectra characteristics are concomitant with the XRD results. The as-deposited coating on foil 1 presents wide phosphate bands at wavenumbers below 1200 cm−1 that correspond to poor crystalline or amorphous calcium phosphate (ACP) material [25]. As the annealing time increases, the absorption peaks corresponding to crystalline phases resolve. However, bands are still wide owing to the superposition of the multiple bands of the different calcium phosphate phases, and perhaps because of the presence of remaining low crystalline material. A peak at 3571 cm−1 appears in the sample annealed for 38 min. This peak is associated with the stretching vibration of the OH− group in HA. Therefore, the coating is incorporating the OH− from the water atmosphere while increasing the crystalline apatite phase during the annealing process. The spectra of the coatings deposited in foil 2 present narrower phosphate bands. This is because coatings are mainly constituted by HA. Again, the annealing process promotes an increase in crystallinity that is reflected in better resolved peaks in the phosphate bands and the apparition of the OH− stretching peak of HA. In this case, this peak already arises after 8 min of annealing, and it grows in absorption as the annealing time advances. The micro-Raman spectra found in the different zones of foil 1 and in foil 2 are shown in Fig. 5. In foil 1, the central zone (zone I) of the as-deposited coating presents a broad

FIGURE 4 Infrared spectra of (a) the as-deposited coatings, and the annealed coatings for (b) 8, (c) 23, and (d) 38 min, corresponding to the two foils

peak centred at 950 cm−1 which is associated with ACP [26]. In the samples annealed for 23 and 38 min, this peak appear widened because an additional contribution of a peak at 962 cm−1 which is associated with HA. Therefore, a partial crystallisation of the amorphous phase is taking place during annealing. Around the central zone in the elliptical ring of zone II, the Raman signal gives two peaks at 947 cm−1 and 968 cm−1 which are characteristic of TCP. The spectra in this zone do not evidence important changes with the annealing time. Outside the ring but still in foil 1 (zone III), the spectrum of the as-deposited coating also presents a broad peak slightly shifted to high Raman shifts with respect to the ACP peak found in zone I. Again, this is because an additional contribution of a small peak around 962 cm−1 . Therefore, the spectrum observed indicates that the material here is not completely amorphous but has some HA. For the annealed coatings, the narrow peak at 962 cm−1 corresponding to HA grows while the broad peak at 950 cm−1 corresponding to ACP diminishes. In foil 2, the spectrum of the as-deposited coating also has the broad peak of ACP and the narrow peak of HA. The spectra of the annealed coatings show a similar evolution than those of zone III, but in all cases the ACP contribution is lower and it already disappears in the sample annealed for 23 min. Moreover, a shoulder at higher Raman shifts appears. The position of this shoulder coincides with the most intense peak of TCP indicating the presence of this phase that is also detected by XRD.

254

Applied Physics A – Materials Science & Processing

FIGURE 5 Micro-Raman spectra at different zones of (a) the as-deposited coatings, and the annealed coatings for (b) 8, (c) 23, and (d) 38 min, corresponding to the three zones of foil 1 and to foil 2

4

Discussion

The composition and structure of a coating on a substrate are determined by different factors during its growth; in particular, the amount of particles arriving to the substrate, their energy, the ratios between the different type of particles and the substrate temperature. In a PLD process, the amount and energy of particles do not have an homogeneous spatial distribution in the plume, and the particlesratios, which we refer as plume composition, could very probably be inhomogeneous as well. Therefore, inhomogeneities in a PLD coating can be produced by inhomogeneities of any of these factors. The variations in the phases observed in this work could be directly related to inhomogeneities of the plume composition, as often considered in PLD coatings [27, 28]. However, because the composition distribution of the plume is unknown and other factors can account at the same time, it is not possible to unequivocally attribute the inhomogeneities in the coatings to it. On the other hand, the presence of amorphous material is a common consequence of a high amount of particles arriving to the substrate. Moreover, changes in the deposition rate, which is basically determined

by this amount of particles, were already correlated with the presence of different calcium phosphate phases in PLD coatings [22]. Furthermore, coatings composed of only HA were obtained under very similar conditions of deposition, but at a lower deposition rate [23]. In these coatings, the unique differences in the technological parameters of deposition are that the amount of energy of each laser pulse and the area of the laser spot are half the energy and half the area used in the present work. Thus, the deposition rate is significantly different, while the laser fluence, a highly determinant factor of the ablation process, is equal in both cases. Hence, as the fluence does not change, the ablation process should not be very different in both cases. In consequence, the main differences between the coatings of both works could be more related to the deposition rate than to plume composition and particle energy effects. With the hypothesis that the main factor affecting the coating characteristics obtained in this work is the deposition rate, many aspects of the phase distribution can be reasonably explained as a consequence of the gradient of deposition rates resulting from the directionality of PLD. Indeed, it was found that high deposition rates (around 0.1 nm/pulse) favoured the

´ FERN ANDEZ - PRADAS et al.

Inhomogeneity of calcium phosphate coatings deposited by laser ablation

presence of α-TCP in HA coatings deposited with a wavelength of 355 nm [22]. High deposition rates do not allow all the material to react with the water of the atmosphere and non-hydroxylated phases like α-TCP are formed. The minor deposition rate studied in the present work is 0.2 nm/pulse occurring in foil 2. There, a composition of HA plus α-TCP is also found, but with an additional contribution of ACP. As the deposition rate is still too high, all the material arriving at the substrate cannot order to form a crystalline structure. This explanation is concomitant with the fact that the amount of ACP is relatively higher in zone III of foil 1 than in foil 2, because deposition is also higher. In zone I, where the deposition rate is the highest (0.5 nm/pulse), only ACP is formed. The amount of material arriving at the substrate is too high, and cannot form a crystalline structure. Only zone II supposes an abrupt change in the progressive disappearance of crystalline phases from foil 2 to zone I as the deposition rate increases. The deposition rate in this zone is about 0.45 nm/pulse, not much lower than in zone I. The TCP found by Raman in zone II probably corresponds to all the β-TCP phase detected by XRD in foil 1 and some α-TCP. This is because the β-TCP phase is only detected in foil 1, while the α-TCP phase is present in both foils. Then, as foil 2 seems to be an extension of zone III, probably this zone is constituted by α-TCP together with ACP and HA. Therefore, the β-TCP would be only present in zone II. Nevertheless, the cause for this presence of β-TCP remains without a clear and contrasted explanation. Surely, the effects involved are related to the bombardment of the coating with the energetic ionic and neutral particles of the ablation plume that arrive at the substrate. These effects would be superimposed onto the effect of the deposition rate. During the process of annealing under the same deposition conditions of temperature and atmosphere, it has been seen that the coating incorporates OH− groups from the water atmosphere, and it progressively transforms from ACP into crystalline phases. This is an indication that these processes cannot take place rapidly enough during deposition owing to the fast deposition rate, and therefore they continue after deposition during annealing. The different calcium phosphate phases behave differently when implanted in the body. HA is the only stable phase, while the others have different degrees of resorption. For this reason, it is important to control the homogeneity of the coatings so as to obtain an homogeneous response of the implant. While the movement of the surface during deposition can solve the inhomogeneity of the coating thickness, this process could not avoid the presence of undesired phases owing to the inhomogeneity of deposition rate intrinsic to PLD. To obtain homogeneous coatings of only HA that could be bioactive and stable, the deposition rate should be slow enough to permit all the material to react and/or order. As mentioned above, HA coatings can be obtained under the same conditions of deposition of the present work, included the fluence, except for a lower laser energy per pulse [23]. This decrease in energy leads to a decrease of the deposition rate. Then, HA is obtained at the centre of the coating, where the deposition rate is the fastest, and also at the rest of the coating, where the deposition rate is slower. Therefore, the deposition rate is a key parameter that cannot be disregarded in the deposition of calcium phosphate

255

coatings by PLD. While the laser fluence is the essential parameter to determine the conditions of ablation, the energy delivered in each laser pulse is also a significant factor in the process of deposition, as it directly affects the deposition rate. The sole indication of the laser fluence found in some papers on PLD of HA is not sufficient. The energy delivered by the laser in each pulse cannot be missed. Moreover, the deposition rate should be always mentioned to make a complete description of the experimental process of deposition. 5

Conclusion

Pulsed laser deposition of calcium phosphate coatings from an HA target at high deposition rate produces an inhomogeneous distribution of phases in parallel with a variation of the deposition rate on the substrate. Despite the coating characteristics could depend upon the plume composition distribution and other factors, their main features can be explained by considering that the more important influence is that of the deposition rate. The central part of the coating where deposition rate was 0.5 nm/pulse is composed of ACP. At distant sites from the centre, the deposition rates are lower and crystalline phases (HA and α-TCP) are formed besides ACP. The relative amount of ACP in front of the crystalline phases diminishes as the distance from the centre increases. However, the zone with a minor deposition rate studied in this work (0.2 nm/pulse) is not free of ACP in an as-deposited coating. Therefore 0.2 nm/pulse is still a high deposition rate to obtain crystalline coatings free of amorphous phases. Only an elliptical ring 1 mm wide with crystalline domains of β-TCP near the centre of deposition breaks the progressive increase of ACP with the deposition rate in the coatings. During annealing under the same deposition conditions of temperature and atmosphere, the ACP material transforms progressively into crystalline phases. Nevertheless, coatings obtained in a previous work under the same conditions of temperature, pressure and laser fluence but at lower deposition rates of around 0.09 nm/pulse were composed of only HA. In consequence, it seems clear that such low deposition rates are required to obtain homogeneous coatings of pure HA that could offer stability at the same time as bioactivity to an implant. ACKNOWLEDGEMENTS This work is part of a research program financed by MCYT of the Spanish Government (Project MAT980334-C02-01) and DURSI of the Catalan Government (1999SGR00056). The authors acknowledge the Serveis Cient´ıfico-T`ecnics of the Universitat de Barcelona for the use of its facilities and the assistance of its technicians.

REFERENCES 1 A.S. Posner: Clin. Orthop. 200, 87 (1985) 2 H.W. Denissen, K. de Groot, J. Eestermans, P.J. Klopper, P.C. Makkes, A. van den Hooff: Ultramicroscopy 5, 124 (1980) 3 C.P.A.T. Klein, J.M.A. de Blieck-Hogervorst, J.G.C. Wolke, K. de Groot: Biomaterials 11, 509 (1990) 4 F.L.J.A. de Wijs, C. de Putter, G.L. de Lange, K. de Groot: J. Prosthet. Dent. 69, 514 (1993) 5 W. Suchanek, M. Yoshimura: J. Mater. Res. 13, 94 (1998) 6 P. Baeri, L. Torrisi, N. Marino, G. Foti: Appl. Surf. Sci. 54, 210 (1992) 7 C.M. Cotell: Appl. Surf. Sci. 69, 140 (1993)

256

Applied Physics A – Materials Science & Processing

8 G. Sardin, M. Varela, J.L. Morenza: Hydroxyapatite and Related Materials, ed. by P.W. Brown, B. Constanz, (CRC Press, London 1994) p. 225 9 R.K. Singh, F. Qian, V. Nagabushnam, R. Damodaran, B.M. Moudgil: Biomaterials 15, 522 (1994) 10 M. Jel´ınek, V. Olˇsan, L. Jastrab´ık, V. Studnicka, V. Hnatowicz, J. Kv´ıtek, V. Havr´anek, T. Dost´alov´a, I. Zergioti, A. Petrakis, E. Hontzopoulos, C. Fotakis: Thin Solid Films 257, 125 (1995) 11 E.N. Antonov, V.N. Bagratashvili, V.K. Popov, E.N. Sobol, M.C. Davies, S.J.B. Tendler, C.J. Roberts, S.M. Howdle: Biomaterials 18, 1043 (1997) 12 C.K. Wang, J.H. Chern Lin, C.P. Ju, R.P.H. Chang: Biomaterials 18, 1331 (1997) 13 J.M. Fern´andez-Pradas, G. Sardin, L. Clèries, P. Serra, C. Ferrater, J.L. Morenza: Thin Solid Films 317, 393 (1998) 14 J.L. Arias, F.J. Garc´ıa-Sanz, M.B. Mayor, S. Chiussi, J. Pou, B. Le´on, M. P´erez-Amor: Biomaterials 19, 883 (1998) 15 H. Zeng, W.R. Lacefield, S. Mirov: J. Biomed. Mater. Res. 50, 248 (2000) 16 L. Clèries, J.M. Fern´andez-Pradas, G. Sardin, J.L. Morenza: Biomaterials 19,1483 (1998)

17 L. Clèries, J.M. Fern´andez-Pradas, J.L. Morenza: J. Biomed. Mater. Res. 49, 43 (2000) 18 L. Clèries, J.M. Fern´andez-Pradas, J.L. Morenza: Biomaterials 21, 1861 (2000) 19 P. Serra, J.L. Morenza: Appl. Phys. A 67, 289 (1998) 20 P. Serra, J.L. Morenza: Thin Solid Films 345, 43 (1998) 21 K.L. Saenger: Pulsed Laser Deposition of Thin Films, ed. by D.B. Chrisey, G.K. Hubler (Wiley, New York 1994) p. 199 22 J.M. Fern´andez-Pradas, L. Clèries, G. Sardin, J.L. Morenza: J. Mater. Res. 14, 4715 (1999) 23 J.M. Fern´andez-Pradas, L. Clèries, E. Mart´ınez, G. Sardin, J. Esteve, J.L. Morenza: Biomaterials 22, 2171 (2001) 24 J.M. Fern´andez-Pradas, L. Clèries, P. Serra, G. Sardin, J.L. Morenza: Appl. Phys. A 72, 613 (2001) 25 C.B. Baddiel, E.E. Berry: Spectrochim. Acta 22, 1407 (1966) 26 M. Weinlaender, J. Beumer III, E.B. Kenney, P.K. Moy: J. Mater. Sci.: Mater. Med. 3, 397 (1992) 27 V.N. Bagratashvili, E.N. Antonov, E.N. Sobol, V.K. Popov, S.M. Howdle: Appl. Phys. Lett. 66, 2451 (1995) 28 J.L. Arias: PhD Thesis (Universidade de Vigo, 2000)