**3.2 Bioactivity tests: Observations & considerations**

*Figure 7* displays a comparison of the BG films IR spectra before and after immersion in simulated body fluid up to 30 days.

For the as-deposited and annealed BG1 structures, the FTIR measurements (*Fig. 7-a,b*) showed no changes in intensity and position of the original vibration bands with immersion time, suggesting the inert character of this material.

In case of as-deposited BG2 samples (*Fig. 7-c*) after only 24 hours of immersion, dramatic changes were observed. The spectrum revealed the disappearance of the initial BG film vibration bands and the emergence of two new and strong bands positioned at 816 and 1107 cm-1, which could be assigned to bending and stretching bands of the amorphous silica phase (Bal et al., 2001). Weaker bands at 736, 892 and 968 cm-1, were noticed, belonging most probably to the original BG layer. The intensity of these bands decreased with the immersion time suggesting a continuously leaching of BG ions into the SBF solution. After 30 days of immersion the BG films is completely dissolved, no IR bands could be emphasized, indicating the resorbability of this material.

The FTIR results of the annealed BG2 samples after in-vitro testing in SBF are presented in *Figure 3-d*. After 24 hours soaking in SBF solution the amplitude of the peak at 924 cm-1 present increased values and displaced at ~ 940 cm-1. This maximum peak at 940 cm-1 is

Magnetron Sputtered BG Thin Films: An Alternative Biofunctionalization

known to have better bioactivity and osteoinductivity (Spence et al., 2008).

HA.

The GIXRD patterns are presented in *Fig. 8* revealed after 30 days immersion in SBF the characteristic lines of hydroxyapatite (HA) as large, overlapping peaks (ICDD: 9–432 , for instance) for all the samples. There are deviations of relative intensities with respect to the reference ICDD card, due probably to imperfect stoichiometry. The increased relative intensity of the (002) line is often reported for chemically grown HA layers and it is assigned to the preferred orientation of crystallites and is typical for the biological apatite. The HA peaks are broad due either to the small crystallite size or to lattice disorder or to both (G.E. Stan et al., 2011). The mean crystallite size along the c axis of the hexagonal HA structure, estimated from the (002) line broadening by using the Scherrer formula, is about 25-30 nm (strain-broadening neglected). Because of the large line broadening one can not distinguish between different HA types, for instance between HA and carbonated

After three days of immersion in SBF a strong and narrow diffraction peak appeared in the GIXRD patterns of BG3-1, BG3-2 and BG3-3 at 2θ = 26.5°, whose intensity is correlated with that of a weak line at 2θ = 54.6°. The most plausible phase is calcium phosphate hydrate type (i.e. Ca(H2P2O7) ICDD: 70-6384), whose role in the subsequent HA formation has been the

subject of several detailed studies (Liu et al., 2008).

Approach – Peculiarities of Bioglass Sputtering and Bioactivity Behaviour 87

the spectrum and the lower contribution of silicates bands overlapped by the emergence of dominant phosphate bands due to Ca–P precipitation. The leaching of BG ions along with the dissolution-diffusion of silicon atoms from the glass structure resulting in a supersaturation at the film-SBF interface is a prerequisite condition for the starting of precipitation process (Hench, 1991). One can hypothesize that the peak positioned at ~1125 cm-1 could obscure also the presence of Si-O-Si asymmetric stretching vibrations (Hench, 1991; Socrates, 2001), owned to the silica-rich layer formed at the film-solution interface at an early stage (Hench, 1991). After 15 days (*Fig. 7-f*), all the BG3 samples displayed similar IR spectra, defined by the phosphate's two intense shoulders positioned around 1028 and 1121 cm-1, respectively. Increases in intensity of water and O–H bands along with the appearance of (CO3)2- stretching vibration bands suggest a continuous incorporation of various molecular structures at the liquid-solid interface. After 30 days of immersion the FTIR absorbance spectra (*Fig. 7-g*) revealed strong vibrations at the following wave numbers: 876, 960, 1019, 1107, 1413, 1465, and 1635 cm–1, corresponding to crystalline carbonated apatite (CHA). The sharp bands at 1019 cm-1 and 1107 cm-1 correspond to (ν3) asymmetric stretching of phosphate groups. The splitting of the stretching and bending IR absorption bands in two narrow components suggests a crystalline apatitic growth. The weak band centred at 960 cm-1 is assigned to (PO4)3- (ν1) symmetric stretching mode. All apatitic layers obtained were hydroxylated and carbonated as demonstrated by the presence of strong O–H bands (the bending mode centred at ~1635 cm-1) and the sharp C–O bending (ν2) and stretching (ν3) lines at (876, 1413 and 1465 cm-1). Considering the hydroxyapatite structure, the carbonate group can substitute both the hydroxyl and the phosphate ions, giving rise to the A-type and B-type carbonation, respectively. The positions of the carbonate bands indicate that (CO3)2 groups are substituting (PO4)3- ions, suggesting the predominance of B-type CHA (Markovic et al, 2004). The B-type is the preferential substitution in the human bone and is

also present related to the Si-OH groups (Berbecaru et al., 2010). In comparison the peak positioned at 1030 cm-1 is decreased in amplitude. The process could be related to the continuous leaching of Na, Ca and Si ions in the SBF solution. The exchange of the ions between the surface layer and SBF solution lead to a rearranging of the bonds in the sample surfaces. At the same time the diffusion of the H+ ions, because of the increase of the electronegativity of the surface toward the BG films will initiate the formation of the Si-OH bonds on their surface (L.L. Hench & J. Wilson, 2003; Berbecaru et al., 2010). The small shoulder at 895 cm-1 revealed the presence of the stretching vibrations of the Si-O bonds with two non-bridging oxygens (Q1 and Q0 units) from the original BG layer. After 72 hours of immersion in SBF solution the peaks at 776 and 1030 cm-1 increased in amplitude suggesting the enrichment of the Si-O-Si bonds in the SiO4 tetrahedra (Q4 units). This suggests that after a certain period of time the polymerization reaction begins at the film surface. Also, the peak at 940 cm-1 strongly decreases in amplitude, revealing three new weak bands at 960 cm-1, 1021 and 1087 cm-1 respectively. These absorption bands are assigned to ν1 symmetric stretching mode of (PO4)3- (960 cm-1) and to ν<sup>3</sup> asymmetric stretching of the phosphate groups, respectively (1021 and 1087 cm-1 bands) (Socrates, 2001).

This phenomena could be related to simultaneous phenomena such: formation of the silicic acid Si(OH)4 and subsequent polycondensation reaction, along with the beginning of the precipitation of calcium phosphate (Ca-P) phases on the surface from the SBF solution suprasaturated in Ca-P ions (Berbecaru et al., 2010). The released water in the polycondensation reaction is known to remain physically bonded with the Si-O-Si surface forming the hydrated silica rich layer (Hench, 1991; L.L. Hench & J. Wilson, 2003). This is leading to an increase of the pH at the surface level which will be favourable to the absorption of the cations and anions on the surface, the precipitation processes of Ca-P rich phases being under these auspicious conditions.

The process of precipitation seems to continue up to 30 days, the relative integral area of the bands at 960 cm-1, 1021 and 1087 cm-1 monotonously increasing with immersion time. Thus, it is suggested that on top of the annealed BG1 coatings chemically develops in vitro a Ca-P type layer. The broad aspect of the bands suggests the amorphous nature of this chemically grown layer. For comparison the spectrum of a crystalline synthetic hydroxyapatite powder is displayed. No crystallization processes were observed in XRD after 30 days of immersion.

*Figure 7-e,f,g* displays the evolution of the FTIR spectra for the BG3 structures after SBF immersions for 3, 15 and 30 days. Similar IR spectral evolutions were observed for all BG3 films. A change of the BG3 spectra envelope was noticed after 3 days of immersion in SBF (*Fig. 7-e*). There are some perceptible differences in shape and amplitude of the IR spectra between these three investigated glass films. The two dominant maxima (~1090 and ~1120 cm-1, respectively) are assigned to the (PO4)3- (ν3) asymmetric stretching, hinting that after three days of immersion the precipitation of a Ca–P layer had started (G.E. Stan et al., 2011). A splitting of the phosphate stretching band, more clearly in case of BG3-4, could be observed, suggesting a more advanced stage of amorphous calcium phosphates phases' precipitation. Thus, at this point, one can deduce that partial dissolution of BG3 structures occurred up to 3 days, as demonstrated by the diminishing overall intensity of

also present related to the Si-OH groups (Berbecaru et al., 2010). In comparison the peak positioned at 1030 cm-1 is decreased in amplitude. The process could be related to the continuous leaching of Na, Ca and Si ions in the SBF solution. The exchange of the ions between the surface layer and SBF solution lead to a rearranging of the bonds in the sample surfaces. At the same time the diffusion of the H+ ions, because of the increase of the electronegativity of the surface toward the BG films will initiate the formation of the Si-OH bonds on their surface (L.L. Hench & J. Wilson, 2003; Berbecaru et al., 2010). The small shoulder at 895 cm-1 revealed the presence of the stretching vibrations of the Si-O bonds with two non-bridging oxygens (Q1 and Q0 units) from the original BG layer. After 72 hours of immersion in SBF solution the peaks at 776 and 1030 cm-1 increased in amplitude suggesting the enrichment of the Si-O-Si bonds in the SiO4 tetrahedra (Q4 units). This suggests that after a certain period of time the polymerization reaction begins at the film surface. Also, the peak at 940 cm-1 strongly decreases in amplitude, revealing three new weak bands at 960 cm-1, 1021 and 1087 cm-1 respectively. These absorption bands are assigned to ν1 symmetric stretching mode of (PO4)3- (960 cm-1) and to ν<sup>3</sup> asymmetric stretching of the phosphate groups, respectively (1021 and 1087 cm-1 bands)

This phenomena could be related to simultaneous phenomena such: formation of the silicic acid Si(OH)4 and subsequent polycondensation reaction, along with the beginning of the precipitation of calcium phosphate (Ca-P) phases on the surface from the SBF solution suprasaturated in Ca-P ions (Berbecaru et al., 2010). The released water in the polycondensation reaction is known to remain physically bonded with the Si-O-Si surface forming the hydrated silica rich layer (Hench, 1991; L.L. Hench & J. Wilson, 2003). This is leading to an increase of the pH at the surface level which will be favourable to the absorption of the cations and anions on the surface, the precipitation processes of Ca-P rich

The process of precipitation seems to continue up to 30 days, the relative integral area of the bands at 960 cm-1, 1021 and 1087 cm-1 monotonously increasing with immersion time. Thus, it is suggested that on top of the annealed BG1 coatings chemically develops in vitro a Ca-P type layer. The broad aspect of the bands suggests the amorphous nature of this chemically grown layer. For comparison the spectrum of a crystalline synthetic hydroxyapatite powder is displayed. No crystallization processes were observed in XRD

*Figure 7-e,f,g* displays the evolution of the FTIR spectra for the BG3 structures after SBF immersions for 3, 15 and 30 days. Similar IR spectral evolutions were observed for all BG3 films. A change of the BG3 spectra envelope was noticed after 3 days of immersion in SBF (*Fig. 7-e*). There are some perceptible differences in shape and amplitude of the IR spectra between these three investigated glass films. The two dominant maxima (~1090 and ~1120 cm-1, respectively) are assigned to the (PO4)3- (ν3) asymmetric stretching, hinting that after three days of immersion the precipitation of a Ca–P layer had started (G.E. Stan et al., 2011). A splitting of the phosphate stretching band, more clearly in case of BG3-4, could be observed, suggesting a more advanced stage of amorphous calcium phosphates phases' precipitation. Thus, at this point, one can deduce that partial dissolution of BG3 structures occurred up to 3 days, as demonstrated by the diminishing overall intensity of

(Socrates, 2001).

phases being under these auspicious conditions.

after 30 days of immersion.

the spectrum and the lower contribution of silicates bands overlapped by the emergence of dominant phosphate bands due to Ca–P precipitation. The leaching of BG ions along with the dissolution-diffusion of silicon atoms from the glass structure resulting in a supersaturation at the film-SBF interface is a prerequisite condition for the starting of precipitation process (Hench, 1991). One can hypothesize that the peak positioned at ~1125 cm-1 could obscure also the presence of Si-O-Si asymmetric stretching vibrations (Hench, 1991; Socrates, 2001), owned to the silica-rich layer formed at the film-solution interface at an early stage (Hench, 1991). After 15 days (*Fig. 7-f*), all the BG3 samples displayed similar IR spectra, defined by the phosphate's two intense shoulders positioned around 1028 and 1121 cm-1, respectively. Increases in intensity of water and O–H bands along with the appearance of (CO3)2- stretching vibration bands suggest a continuous incorporation of various molecular structures at the liquid-solid interface. After 30 days of immersion the FTIR absorbance spectra (*Fig. 7-g*) revealed strong vibrations at the following wave numbers: 876, 960, 1019, 1107, 1413, 1465, and 1635 cm–1, corresponding to crystalline carbonated apatite (CHA). The sharp bands at 1019 cm-1 and 1107 cm-1 correspond to (ν3) asymmetric stretching of phosphate groups. The splitting of the stretching and bending IR absorption bands in two narrow components suggests a crystalline apatitic growth. The weak band centred at 960 cm-1 is assigned to (PO4)3- (ν1) symmetric stretching mode. All apatitic layers obtained were hydroxylated and carbonated as demonstrated by the presence of strong O–H bands (the bending mode centred at ~1635 cm-1) and the sharp C–O bending (ν2) and stretching (ν3) lines at (876, 1413 and 1465 cm-1). Considering the hydroxyapatite structure, the carbonate group can substitute both the hydroxyl and the phosphate ions, giving rise to the A-type and B-type carbonation, respectively. The positions of the carbonate bands indicate that (CO3)2 groups are substituting (PO4)3- ions, suggesting the predominance of B-type CHA (Markovic et al, 2004). The B-type is the preferential substitution in the human bone and is known to have better bioactivity and osteoinductivity (Spence et al., 2008).

The GIXRD patterns are presented in *Fig. 8* revealed after 30 days immersion in SBF the characteristic lines of hydroxyapatite (HA) as large, overlapping peaks (ICDD: 9–432 , for instance) for all the samples. There are deviations of relative intensities with respect to the reference ICDD card, due probably to imperfect stoichiometry. The increased relative intensity of the (002) line is often reported for chemically grown HA layers and it is assigned to the preferred orientation of crystallites and is typical for the biological apatite. The HA peaks are broad due either to the small crystallite size or to lattice disorder or to both (G.E. Stan et al., 2011). The mean crystallite size along the c axis of the hexagonal HA structure, estimated from the (002) line broadening by using the Scherrer formula, is about 25-30 nm (strain-broadening neglected). Because of the large line broadening one can not distinguish between different HA types, for instance between HA and carbonated HA.

After three days of immersion in SBF a strong and narrow diffraction peak appeared in the GIXRD patterns of BG3-1, BG3-2 and BG3-3 at 2θ = 26.5°, whose intensity is correlated with that of a weak line at 2θ = 54.6°. The most plausible phase is calcium phosphate hydrate type (i.e. Ca(H2P2O7) ICDD: 70-6384), whose role in the subsequent HA formation has been the subject of several detailed studies (Liu et al., 2008).

Magnetron Sputtered BG Thin Films: An Alternative Biofunctionalization

Approach – Peculiarities of Bioglass Sputtering and Bioactivity Behaviour 89

Fig. 8. Comparative representation of the GIXRD patterns of the as-deposited BG films before and after 3, 15 and 30 days immersion in SBF. = Hydroxylapatite (ICDD:9–432);

The intensity of the line at 2θ = 26.5° start to diminish until the 15th day of immersion in SBF, and almost disappears after 30 days for all the samples. It is interesting to note that this phase has a strong intensity only in the structures deposited in pure Ar (BG3-1, BG3-2 and BG3-3) and it is very weak for the BG3-4 samples deposited in reactive atmosphere (7% oxygen). Another quite intensive line in the GIXRD patterns of the samples after SBF immersion appears at 2θ = 36.0°. The best assignment for this line is TiH (ICDD: 40-1244). This peak is present in all the structures that were immersed in SBF and it is expected that

= TiH1.7 (ICDD:40–1244); ▲= calcium phosphate hydrate (ICDD: 70-6384)

its presence is related to the uncovered regions of the titanium substrate.

Fig. 7. Evolution of FTIR spectra of the BG films with increasing immersion time in SBF (1 - 30 days)

Fig. 7. Evolution of FTIR spectra of the BG films with increasing immersion time in SBF (1 -

30 days)

Fig. 8. Comparative representation of the GIXRD patterns of the as-deposited BG films before and after 3, 15 and 30 days immersion in SBF. = Hydroxylapatite (ICDD:9–432); = TiH1.7 (ICDD:40–1244); ▲= calcium phosphate hydrate (ICDD: 70-6384)

The intensity of the line at 2θ = 26.5° start to diminish until the 15th day of immersion in SBF, and almost disappears after 30 days for all the samples. It is interesting to note that this phase has a strong intensity only in the structures deposited in pure Ar (BG3-1, BG3-2 and BG3-3) and it is very weak for the BG3-4 samples deposited in reactive atmosphere (7% oxygen). Another quite intensive line in the GIXRD patterns of the samples after SBF immersion appears at 2θ = 36.0°. The best assignment for this line is TiH (ICDD: 40-1244). This peak is present in all the structures that were immersed in SBF and it is expected that its presence is related to the uncovered regions of the titanium substrate.

Magnetron Sputtered BG Thin Films: An Alternative Biofunctionalization

crystallinity could play important roles in BG films' reactivity in SBF.

time.

Approach – Peculiarities of Bioglass Sputtering and Bioactivity Behaviour 91

formation of a non-adherent fibrous interfacial capsule. No surface changes during SBF tests were evidenced for the annealed BG1 structure. Unlike in case of BG2, the crystallization had no effect on BG1's surface reactivity. The nucleated Na2Mg(PO3)4 phase proved to be inert in SBF. It can be concluded that the chemical composition, bonding configuration and

In case of as-deposited BG2 samples a continuous dissolution process occurs, after 30 days the entire film is dissolved into the surrounding fluid. One can notice the high content of Ca (~20 at.%) and Na atoms (~43 at.%) in the BG2 film (*Table 3*). It can be speculated that when very high numbers of alkali and alkali-earth ions are released in SBF, the pH will rapidly and drastically change at the film-fluid interface, producing a chemical unbalance which will suppress the polymerization of silanols and the further formation of the SiO2-rich surface layer. The bioactive process is thus disrupted, leading to continuous dissolution of the BG amorphous film. On the other hand the annealing of BG1 films followed by slow cooling promotes formation of combeite, wollastonite and Na3PO4, which have smaller solubility, and consequently a decreased leaching rate of Na+ and Ca2+ into the SBF solution. This will determine the conditions at the surface of the film to be more stable, therefore allowing polymerisation of soluble Si(OH)4 in a SiO2-rich layer which will act as a nucleation site for apatite (L.L. Hench & J. Wilson, 2003). During 30 days of SBF soaking, dissolution-reprecipitation processes took place, the annealed BG2 layer is partially dissolved and finally we obtain a multilayer structure containing a bottom BG layer coated by an amorphous SiO2-rich thin film and at the top a Ca-P type layer. Thus in case of BG2 annealed samples our FTIR results are consistent with Hench's theory. Previous studies have also reported that both the combeite (Chen et al., 2006) and wollastonite (Xue et al., 2005) can generate on their surface Ca/P rich layers when in contact with SBF. The growth of the bioapatite layer is essential for bone generation and bonding ability. However, the combeite was proved to rapidly transform into amorphous calcium phosphate phase in contact with SBF, but it delays the process of crystallization into a hydroxyapatite phase (L.L. Hench & J. Wilson, 2003). Besides biocompatibility, these structures could actively improve proteins and osteocytes adhesion, significantly shortening the osteointegration

Regarding the higher biomineralization rates of BG3-3 and BG3-4 films one can hypothesize that the chemical processes involved in the bioactivity mechanism are accelerated, because of the increased sodium content of these films and a propitious bridging oxygen/nonbridging oxygen ratio (*Table 4*), speeding up the ionic exchange and the chemical growth of HA. Hench's theory states that the first stage of SBF immersion of a bioglass involves the rapid exchange of Na+ ions from the glass for H+ and H3O+ ions from the solution, which shall initialize the formation of silica-rich layer known to favour the nucleation of Ca-P type layers. Therefore, we can suppose that a higher number of sodium ions released into SBF solution will change the chemical equilibrium of the precipitation reaction, thereby

Therefore, for achieving bioactive properties, an optimal ratio of network formers/network modifiers is required, but not sufficient. The disruption of the Si-O-Si bonds and creation of Si-O-NBO bonds, determined by increased network modifiers content, is mandatory in the first steps of Ca-P chemical growth mechanism. The disruption of the Si-O-Si bonds evidenced by the increasing content of Si-O-NBO groups plays an important role in the

catalyzing the CHA chemical growth mechanism.

Fig. 9. SEM micrographs of BG3 films after immersion in SBF for 30 days. Left side, SEM-top view; Right side, ESEM-cross view

The SEM microstructures of BG3 films immersed for 30 days clearly asserted the growth of thick and rough coatings with randomly distributed irregular spherulitic shaped microsized agglomerates on top of all three types of samples, evidencing the good biomineralization capability of all coatings (*Fig. 9-top row*). The thicknesses of the chemical grown HA layers were determined by tilt-SEM (*Fig. 9-bottom row*). One can observe that the BG3-3 and BG3-4 coatings led to the thickest chemical growths: total film thickness ~1 μm, and ~1.3 μm, respectively, compared to the as-deposited film thickness of ~ 0.4 μm. The HA chemically grown layers had in case of BG3-1 and BG3-2 structures a lower average thickness of ~0.7 μm. Thus, the best biomineralization, correlated with the thickest HA layer growth, was obtained for the BG3-4 coating (G.E. Stan et al., 2011).
