The influence of cobalt incorporation and cobalt precursor selection on the structure and bioactivity of sol-gel-derived bioactive glass

Breno R. Barrioni1,*, Elizabeth Norris2, Julian R. Jones2 and Marivalda de M. Pereira1

1Department of Metallurgical Engineering and Materials, Federal University of Minas Gerais, School of Engineering, Belo Horizonte, MG, Brazil

2Department of Materials, Imperial College London, South Kensington Campus, London, SW7 2AZ, UK

*Corresponding author. Tel.: +5531984764001; E-mail: brenorb@ufmg.br (B. R. Barrioni)

E-mail addresses: brenorb@ufmg.br (B. R. Barrioni), e.norris14@imperial.ac.uk (E. Norris), julian.r.jones@imperial.ac.uk (J. R. Jones), mpereira@demet.ufmg.br (M. M. Pereira)

Abstract

Cobalt (Co) is a potential therapeutic ion used to enhance angiogenesis through a stabilizing effect on hypoxia-inducible factor 1 alpha (HIF-1α), and its incorporation into the structure of bioactive glass is a promising strategy to enable sustained local delivery of Co to a wound site or bone defect. Here, Co-releasing bioactive glasses were obtained through the sol-gel method, comparing cobalt nitrate and cobalt chloride as precursors. The effect of using different Co precursors on the sol-gel synthesis and in the obtained bioactive glass structure, chemical composition, morphology, dissolution behaviour, hydroxycarbonate apatite (HCA) layer formation was investigated. When the chloride salt was used as Co precursor, evidence of crystalline cobalt (II, III) oxide (Co3O4) phase formation was found, along with the presence of Co3+ species as evaluated by X-ray Diffraction (XRD) and X-ray photoelectron spectroscopy (XPS), whereas an amorphous glass containing mainly Co2+ species was obtained when the nitrate salt was the Co source. The presence of a crystalline phase decreased the surface area and pore volume of the final glass, consequently reducing the Co-release rate. Evidence of HCA layer formation after immersion in simulated body fluid (SBF) was still found when different precursors were used, although the rate of formation was reduced by the presence of Co. Therefore, this study showed that Co incorporation and the proper selection of the precursor could affect the final material structure, and properties, and should be considered when designing new bioactive glasses compositions for tissue engineering applications.

Keywords: bioactive glass; sol-gel; cobalt; bioactivity

1. Introduction

In 1969, Larry Hench observed that a special glass composition derived from the Na2O – CaO – SiO2 – P2O5 system formed a strong connection with the host tissue after being implanted in the femur of rats [1]. This bioactive glass composition, later called Bioglass® or 45S5, became an alternative to inert implant materials, presenting high biocompatibility and bioactivity, inducing rapid bone tissue regeneration [2]. Different glass compositions, such as 70S30C and 58S, have been subject to intense research in the last decades [3,4], founding several applications in tissue engineering, from hard to soft tissue repair [5].

The sol-gel process allows the synthesis of silica-based materials ranging from micro [3] to nanoparticles with different morphologies [6,7], the incorporation of drugs and biomolecules on their structure [8,9], and the synthesis of organic-inorganic hybrid systems [10] of great interest as biomaterials. Sol-gel derived bioactive glasses presents advantages over melt-derived glasses such as a higher surface area and an inherent nanoporosity, resulting in improved cellular responses and higher bioactivity [11–13]. Also, as sol-gel is a versatile process, bioactive glasses can be obtained in several compositional systems, containing different ions with therapeutic properties incorporated into the glass network. During the bioactive glass dissolution process, these ions with physiological activities are released by ion exchange reactions with the surrounding fluid, playing a specific role in the glass biofunctionality. For example, calcium and soluble silica release could favour osteoblast differentiation [14,15], whereas silver and gallium release have shown antibacterial properties [16,17]. Strontium is also known to simulate bone-forming cells while also prevent bone resorption [18,19]. The bioactive glass performance can be improved by the incorporation of these ions into the glass structure, and the glass structure enables the local sustained delivery of these ions as the glass dissolves [20].

Angiogenesis has an important role in the tissue repair, and strategies to induce angiogenesis are being considered when designing materials for regenerative medicine [21]. Transition metals, such as cobalt, have been shown to promote angiogenesis by simulating hypoxia conditions through the stabilization of the Hypoxia Inducible Factor (HIF-1α) [22]. The HIF-1α pathway activates different pro-vasculogenic genes, such as the vascular endothelial growth factor (VEGF), essential for angiogenesis, and consequently improving tissue repair [23,24]. Therefore, Co-releasing glasses have been proposed as an artificial stabilizer of HIF-1α with the consequent promotion of neovascularization for tissue repair. Previous works have shown an increase in the expression of different genes related to the hypoxic response, extracellular matrix remodelling and angiogenesis during in vitro studies in Co-containing bioactive glasses [22,24–27]. However, the controlled release of Co should be obtained to prevent toxicity, as previously observed in a concentration-dependent manner in hMSC (human mesenchymal stem cells) cultured with Co-containing glass dissolution products [24,28].

The precursor selection on sol-gel bioactive glass synthesis can affect its final structure and properties [29]. For example, the use of different phosphorous precursors in previous work (triethyl phosphate, phosphoric acid and phosphorous oxide dissolved in ethanol) resulted in different gelation times and mineralization behaviour on bioactive glasses and glass-ceramics in the SiO2 – P2O5 – CaO composition, although still presenting bioactivity [29,30]. Also the precursor of network modifiers such as calcium was evaluated, and showed that calcium only enters the glass structure above 400 °C when calcium nitrate is used as precursor [31] but enters at the glass network at room temperature when using calcium methoxyethoxide (CME), and it is not incorporated at any temperature when the precursor is the calcium chloride salt [4]. Likewise, the lithium precursor was shown to influence the final glass structure, as the use of lithium nitrate leads to the formation of crystalline lithium metasilicate, resulting in a dense glass-ceramic, whereas when lithium citrate was used as a precursor, a bioactive glass with lithium as a network modifier was obtained [32]. Sol-gel glasses containing Co were already described by Wu et al. [26], which described the synthesis of hypoxia-mimicking mesoporous bioactive glass scaffolds. Nonetheless, the effect of different Co precursors on the sol-gel synthesis of bioactive glasses was not evaluated before, which could influence the glass properties and dissolution behaviour, consequently affecting the material performance.

Here, we synthesized new sol-gel 58S-derived bioactive glass compositions containing Co as a potential strategy for controlled ion release in physiological environments, evaluating the effect of Co incorporation and precursors (cobalt chloride and cobalt nitrate) on the glass properties.

1. Materials and Methods1. Synthesis of bioactive glass microparticles

Bioactive glass samples derived from the 58S composition (60 mol% SiO2, 4 mol% P2O5 and 36 mol% CaO) were obtained through the sol-gel method as described elsewhere [33,34]. All reagents were purchased from Sigma-Aldrich (UK) unless otherwise stated. Briefly, TEOS (tetraethyl orthosilicate 98%) and TEP (triethyl phosphate 99%) were hydrolysed in an acidic medium (2M nitric acid) containing deionized water (molar ratio H2O/TEOS = 12) for 1 h with magnetic stirring. Then, calcium nitrate (Ca(NO3)2.4H2O) was dissolved in the sol and mixed for 1 h. The gel obtained was aged for 72 h in a sealed container, followed by drying with a 10°C temperature increase every 24 h until reaching 120 °C. After that, the samples were thermally treated at 650 °C at a heating rate of 2 °C/min for 3 h, ground and sieved to obtain particles lower than 150 µm. The 58S bioactive glass sample was named BG.

For Co-containing samples, cobalt chloride (CoCl2.4H2O, Acros Organics, Fisher Scientific, UK) and cobalt nitrate (Co(NO3)2.4H2O) were used as Co precursors and incorporated into the bioactive glass structure by partial replacement of the calcium content as shown in Table 1. The respective salt was dissolved in the sol prior to calcium nitrate for each composition. Samples were named 5Co-N (obtained with cobalt nitrate salt) and 5Co-C (obtained with chloride salt).

1. Characterization of bioactive glass samplesChemical composition

The chemical composition of samples was determined by inductively coupled plasma-optical emission spectrometer (ICP-OES) using a Thermo Scientific iCaP 6000series equipment (Watham, MA, USA). Samples were prepared by acid digestion method, in which 50 mg of each sample were carefully mixed in a platinum crucible with 250 mg of anhydrous lithium metaborate (80% w/w) and lithium tetraborate (20% w/w) (Spectroflux 100B, Alfa Aesar, Lancashire, UK), and then fused for 30 min at 1050 °C in a furnace. After cooling, samples were completely dissolved in 2 M nitric acid. The concentrations of Si, Ca, P and Co in each sample were then determined by ICP-OES, and the proportions of SiO2, CaO, P2O5 and CoO were calculated.

Thermal analysis

Thermogravimetry Analysis (TGA) and Differential Scanning Calorimetry (DSC) were performed using a Netzsch STA 449 C (Germany) instrument in air. The evaluation was performed on samples before thermal stabilization, placed on a platinum crucible and heated from 30 to 1000 °C at a heating rate of 10° C/min; an empty platinum crucible was used as a reference.

Structural evaluation

Samples were analysed by Fourier Transform Infrared Spectroscopy (FTIR) using a Thermo Scientific Nicolet iS10 (Waltham, USA) equipment and the Attenuated Total Reflectance Accessory. Spectra were obtained in the range of 400 to 4000 cm-1 (64 scans per spectrum; 4 cm-1 resolution). A Bruker D2 phaser (Germany) desktop was used to obtain X-Ray Diffraction (XRD) data, using 0.02 step size without spinning and a CuKα radiation source. Diffraction data were collected from 2θ ranging from 7 to 70°. X-Ray Photoelectron Spectroscopy (XPS) was performed on a K-Alpha+ system (Thermo Scientific UK) operating at a base pressure of 2 x 10-9 mbar. A monochromated and microfused Al Kα X-ray source (hν = 1486.6 eV) was incorporated, with a 180° double focusing hemispherical analyser containing a 2D detector. The emission current of the X-ray source was 6 mA and the anode bias was 12 kV. To minimize sample charging a flood gun was used, and any remaining small shifts observed in binding energy (BE) due to sample charging were corrected using the C 1s core line at a binding energy of 285.0 eV. Data were collected using X-ray spot size of 400 µm and 200 eV pass energy for survey and 20 eV pass energy for core level scan. All data were evaluated using the Avantage XPS software package. High-resolution core-level scans were performed for Si 2p, O 1s and Co 2p.

Textural properties

Samples were degassed for 24 h at 200 °C prior to measurement. Nitrogen sorption was performed in an Autosorb AS-6 (Quantachrome, UK) multi-station with 40 absorption points 40 desorption points. To evaluate the surface area of samples the Brunauer-Emmett-Teller (BET) method was applied, and absorption data points in the range of 0.01 – 0.30 (P/P0) relative pressure was used, whereas the pore diameter distribution was evaluated by Barret-Joyner-Halenda (BJH) method applied to the desorption curves.

Dissolution Study

Simulated Body Fluid (SBF) solution was prepared as previously described [35,36]. Bioactive glass powders (75 mg) were immersed in SBF solution (50 mL), in an airtight container, and incubated at 37 °C with 120 rpm agitation. After different time points (4 h, 8h, 24 h, 72 h, 1 week and 2 weeks), 1 mL of the bulk solution was collected and then replaced by 1 mL fresh SBF in each container. The collected aliquot was then diluted in nitric acid (2M) with a 10-fold dilution factor, filtered (0.22µm) and analyzed by ICP-OES for the Si, Ca, P and Co ion release in SBF [3]. The experiment was performed in triplicate.The reacted powders were filtered after the 2 weeks study, washed with acetone, dried, and evaluated by FTIR and XRD. The morphology of the samples after immersion in SBF was evaluated by Scanning Electron Microscopy (SEM), performed on a LEO Gemini 1525 microscope (Carl Zeiss SMT, UK) containing field emission column, operating at 5 kV. Samples were sputter coated with chromium prior to the evaluation.

Statistical Analysis

For the chemical analysis, textural properties and dissolution in DMEM, three measurements were performed for each sample, and results are expressed as mean ± standard deviation. Further statistical analysis was performed on each component of the glass composition and also on the results obtained for the textural properties using One-way ANOVA followed by Tukey’s test (GraphPad Prism 6.01), and differences were considered significant for p<0.05.

1. Results and Discussion2. Chemical composition

Table 1 presents the nominal and measured (% mol) chemical composition of sol-gel bioactive glasses after thermal stabilization at 650 °C, measured by the lithium metaborate fusion dissolution process and ICP-OES. Co-containing samples presented similar compositions, with no significant statistical differences. The P2O5 content showed the higher difference between the nominal and measured composition. Variations on the final glass chemical composition may be related to several factors, such as impurity of the precursors, precipitation of species, absorbed moisture, filtration steps and use of a very diluted medium for ICP-OES analysis [37,38]. Nonetheless, Table 1 shows that the Co precursor does not affect the final composition of the bioactive glass.

Table 1: Nominal and measured composition of bioactive glasses (% mol), determined by lithium metaborate fusion and ICP-OES.

Samples

Nominal composition

(% mol)

Measured composition

(% mol)

SiO2

P2O5

CaO

CoO

SiO2

P2O5

CaO

CoO

BG

60.0

4.0

36.0

-

61.3 ± 0.9

3.2 ± 0.6

35.5 ± 0.9

-

5Co-C

60.0

4.0

31.0

5.0

58.1 ± 1.9

3.4 ± 0.5

33.2 ± 2.3

5.3 ± 0.4

5Co-N

60.0

4.0

31.0

5.0

58.9 ± 0.8

3.1 ± 0.4

32.5 ± 0.5

5.5 ± 0.3

2. Thermal analysis

TGA and DSC analysis were performed on samples after synthesis and the drying process to evaluate the influence of Co incorporation on the bioactive glass thermal behaviour and also to determine the ideal heating temperature for these materials. From the TGA (Fig. 1 (a)) three main regions of weight loss can be observed as samples were heated from 30 to 1000 °C. The first stage was observed from 30 to 180 °C, with approximately 10 % mass loss in all samples, mainly related to the removal of residues from the sol-gel process, such as alcohols, and to the endothermic desorption of physically adsorbed water [39]. As the heating process proceeded from 200 to 450 °C the second stage with further 15 – 25 % weight loss related to the release of chemically adsorbed water and loss of remaining organic groups from the sol-gel precursors is observed [29]. Finally, the last stage, from 450 to 600 °C, with additional 10 – 15% weight loss, can be related to nitrate (NO3-) decomposition, present in high content mainly due to the calcium nitrate used as a precursor. It is in accordance with previous works that have shown that calcium only enters in the glass structure as a network modifier ion after nitrate decomposition at temperatures above 400 °C [4,40]. This weight loss stage can also present the contribution of further condensation of hydroxyls on the glass surface. The higher weight loss observed for BG samples may be related to excess of residuals from the sol-gel process on this sample, such as alcohols and remaining organic groups, and also to removal of adsorbed water on the highly reactive BG surface. After that, the weight loss stabilized for the glasses evaluated.

DSC (Fig. 1 (b)) analysis shows a well-defined exothermic transition in the range of 278 – 286 °C, assigned to the carburization of remaining organic sol-gel precursors [29]. An endothermic transition between 490 – 570 °C related to nitrate decomposition is also observed. This transition is observed in a lower intensity for 5Co-C, which may be related to the fact that in this sample the calcium nitrate salt was partially replaced by the cobalt chloride, whereas the reduced temperature at which this transition was observed for 5Co-N can be related to the fact that nitrates of transition metal salts usually present lower degradation temperature than base metal nitrates, as described elsewhere [41]. The exothermic transition observed above 900°C is assigned to the crystallization of the glasses; however, crystalline phases can affect the bioactivity by reducing the possibility of ion exchange reactions with the physiological environment. Based on the thermal evaluation of bioactive glass samples, the heating programme of bioactive glass samples was set to 650°C at a heating rate of 2 °C.min-1 for 3 h, which should promote total nitrate elimination and prevent crystallization of the silica network.

Fig. 1: (a) Thermogravimetric analysis (TGA) and (b) differential scanning calorimetry (DSC) of bioactive glasses.

2. Structural evaluation

Structural evaluation of bioactive glass samples thermally treated at 650 °C was carried out by FTIR (Fig. 2 (a)), and typical bioactive glass absorption bands were observed. No nitrate absorption bands were identified, confirming that the thermal program was adequate for nitrate removal in all samples. The Si-O bending is assigned to the absorption around 580 cm-1 [42] and is usually related to amorphous silica structures [43], whereas the Si-O symmetric stretching is reported at 810 cm-1 [44]. The asymmetric stretching of Si-O-Si bonds is assigned to the absorbance observed around 1000 - 1200 cm-1 [45], with maximum intensity at 1050 cm-1 for all samples. The maximum absorbance for this vibration is usually observed at 1090 cm-1 in SiO2 samples, but it shifts to lower values when network modifiers are present, increasing the number of non-bridging oxygen in the glass structure, indicating a high network modifier content on the evaluated samples [46]. The asymmetric stretching for SiO4 structures containing n bridging oxygen atoms (Qn units) is related to the bands at 1235 cm-1 and 1060 cm-1 [47]. Furthermore, the band between 890 and 975 cm-1 is usually related to SiO4 tetrahedron structures containing non-bridging oxygen [47,48], however, no differences were observed between the samples and further evaluation should be performed to elucidate the role of Co ions on the glass network.

P-O symmetric stretching (1160 cm-1) [37] and P-O bending (500 – 600 cm-1) [46] are also observed, and could also be also associated with orthophosphate units that can form during synthesis [30]. Evidence of Co(III)-O bonds in Co3O4 structures was also found for the 5Co-C sample. The Co3O4 adopts the normal spinel structure with Co2+ species occupying the tetrahedral sites and Co3+ species in the octahedral interstices [49]. Two distinct bands in the FTIR spectra at around 655 and 555 cm-1 are often described in the literature to the presence of Co3+ species in Co3O4 structures [49–51]. The 5Co-C FTIR showed a band with maximum intensity at around 655 cm-1 that could be related to the presence of Co3+ species, although the band at 555 cm-1 may be overlapped by the P-O vibration in the same region.

The FTIR analysis shows the presence of the main absorptions bands related to bioactive glasses in all the samples evaluated. No major differences were observed between the reference bioactive glass and Co-containing samples synthesized using different precursors, indicating that the glass chemical structure was maintained in spite of Co incorporation. Evidence of Co(III)-O bonds was found for the Co-containing bioactive glass obtained using the chloride Co precursor. To confirm the presence of crystalline structures, XRD was performed and is shown in Fig. 2 (b).

Fig. 2: (a) Fourier transform infrared (FTIR) spectra and (b) X-Ray Diffraction (XRD) patterns of bioactive glass samples.

A broad and diffuse halo between 15 and 35° (2θ) was observed in all samples, suggesting the formation of a mainly amorphous silica structure [39], although the higher intensity between 31 – 33° (2θ) for BG indicates an initial arrangement of the material structure, with small apatite phase nucleus formation during heat treatment, as related elsewhere [30]. Peaks observed for 5Co-C match cobalt (II, III) oxide (Co3O4) structures [52], whereas crystalline cobalt oxides were not detected in 5Co-N. Co3O4 is a stable and poorly soluble compound which presents a mixed valence (Co2+ and Co3+) [51], and its presence indicates that Co was not fully incorporated as a network modifier in the glass network when cobalt chloride was used. Although the main amorphous character was maintained after the heat treatment, crystalline regions may result in a lower dissolution rate, consequently affecting the glass bioactivity. Furthermore, the HIF-1α stabilization effect was mainly related to Co2+ species in previous works [28,53], therefore the presence of Co2+ species on the material structure is preferred over Co3+ to enhance angiogenesis in tissue engineering applications.

To confirm and provide further details on the oxidation state of the species, XPS analysis was carried out and Fig. 3 shows the survey scan XPS of bioactive glass samples, and high-resolution XPS spectra (Fig. 4) performed for Si, O and Co. Typical XPS and Auger lines from the glass constituent elements were observed and are identified in the spectra, along with the “adventitious carbon” (C 1s) peak, related to the absorption of hydrocarbon impurities by the glass surface [48]. This peak was used as a reference for the binding energy (BE) calibration, and it was set at 285 eV to correct the sample charging in all samples.

Fig. 3: XPS spectra of bioactive glass samples.

Fig. 4 (a) shows the Si 2p high-resolution XPS spectra, with a maximum peak at 103.5 eV, only slightly shifted to 103.4 eV for 5Co-C, but within the binding energy typical of tetrahedral Si coordinated with four oxygen atoms (SiO44-) [54]. The O 1s XPS spectra, shown in Fig. 4 (b), could be deconvoluted into two different components, at 531.1 eV and 533.0 eV. Although obtaining reliable O 1s data is difficult due to overlapping oxide peaks, and the possibility of hydroxylation and/or carbonation of the material surface [55], the peak with maximum at ~533.0 eV is often assigned to the oxygen in a silica environment, typical of silica glasses, whereas the lower binding energy peak can be related to oxygen in phosphate structure, in the same region of the binding energy observed for O 1s in phosphates of hydroxyapatite [56]. Previous works show that P atoms are mainly in the orthophosphate (PO43-) form in bioactive glasses, charge-balanced by modifier cations, and only small amounts of Si-O-P bonds are observed [30]. No main difference was observed on the Si 2p and O 1s spectra for BG and Co-containing glasses, which indicates that the main silica structure was maintained after Co incorporation.

Co ions in the glass structure can be found in Co2+ or Co3+ state [57], and although XPS is a tool frequently used to identify the chemical state of most elements in the periodic table, it is not a straightforward process for transition metals 2p spectra as they contain complications such as multiplet splitting and shake-up structures that may prove difficult in the identification of the chemical states present [58]. Nonetheless, some features of the Co 2p spectrum can be considered when identifying the Co oxidation states. The Co2+ 2p BE (781 – 783 eV) is usually slightly higher than the BE observed for Co3+ (779 – 780 eV); strong satellite shake-up structures are observed for high-spin Co2+ species; and Co3+ spin-orbit splitting (15.0 – 15.5 eV) is usually lower than Co2+ spin-orbit splitting (15.7 – 16.0 eV) [59,60].

High-resolution scans of the Co 2p peaks are shown in Fig. 4 (c) and multiple peaks are revealed in the range of 772 to 809 eV. The typical 2p peak pairs from spin-orbit splitting (Co 2p3/2 and Co 2p1/2) are observed in the position around 781 and 797 eV [61,62], along with shakeup satellite features observed at their higher binding energy side. This intense shake-up satellite structures observed on both samples are assigned to high-spin Co2+ species [57,62]. However, a reduction in the satellites intensity for 5Co-C could also suggest the presence of Co3+, as these structures are usually weaker in Co 2p spectra when this oxidation state is present [60,61]. The 5Co-C Co 2p peak width and satellite features are also slightly broader, which is another indicator that Co is present in more than one oxidation state or coordination geometry on this sample [61]. Furthermore, the Co 2p3/2 spectra of 5Co-C revealed a high-intensity shoulder at 779 eV that can be related to the presence of Co3+ species, as this peak is usually observed at lower binding energies than Co2+ [58,59]. Finally, the spin-orbit splitting separation of Co 2p3/2 and Co 2p1/2 was 15.8 eV for 5Co-N and slightly lower for 5Co-C (15.5 eV) and indicates the presence of Co in higher oxidation state for 5Co-C [61]. Therefore, XPS evaluation shows that Co is present mainly as Co2+ in the glass structure, but also Co3+ species can be formed when the Co precursor is cobalt chloride, which is in agreement with the XRD evaluation, that showed formation of Co3O4 structures for 5Co-C, which is a compound formed by the mixture of Co2+ and Co3+ species.

Fig. 4: High-resolution XPS spectra of (a) Si 2p (b) O 1s and (c) Co 2p for bioactive glasses samples

2. Textural properties

Nitrogen sorption analyses were performed to evaluate the textural properties of the bioactive glasses. Representative nitrogen adsorption/desorption isotherms and pore size distribution curve derived from the desorption branch (BJH model) are shown in Fig. 5, and the values obtained from BET and BJH algorithm are summarized in Table 2.

Fig. 5: (a) Nitrogen sorption isotherms and (b) pore size distribution (BJH algorithm on desorption branches) of bioactive glass samples.

Type IV isotherms with the presence of hysteresis loop can be identified in all cases, which is characteristic of materials with pore widths ranging from 2 to 50 nm (mesoporous range) [56]. During the sol-gel drying process, the removal of condensation by-products presents within the gel network results in the formation of nanopores, typical sol-gel glasses [31]. Narrow and monomodal pore size distribution can be observed in Fig. 5(b), with pore diameters within the mesoporous range. No statistically significant differences were observed between the mean pore diameter of the bioactive glasses. Samples also presented a high specific surface area, as shown in Table 2, in accordance with values previously reported [37]. A statistically significant decrease on the specific surface area and total pore volume were observed for 5Co-C when compared to BG, which could be attributed to the material crystalline content, as the crystalline phase can obstruct the nitrogen within the glass, consequently affecting their textural properties [32]. The presence of pores within the mesoporous range is favourable in inducing HCA nucleation and crystallization, as the pores could act as nucleation sites. The number of sites available and the size of the pores could control the rate of nucleation and further crystallization of HCA on the glass surface [63], consequently affecting bioactivity, therefore bioactive glasses containing high surface area and mesoporous structure are of great interest for tissue engineering.

Table 2: Summary of the textural characterization obtained performed on bioactive glasses

Samples

Specific Surface Area (m2/g)

Total Pore Volume (cm3/g)

Mean pore diameter (nm)

BG

124.0 ± 11.8

0.365 ± 0.011

7.6 ± 0.4

5Co-N

113.6 ± 8.7

0.297 ± 0.019

6.6 ± 0.6

5Co-C

96.6 ± 14.3

0.261 ± 0.44

6.4 ± 0.2

2. Dissolution Study

After immersion in SBF, the dissolution process of the bioactive glass is initiated, altering the composition and pH of the solution, providing sites that favour the HCA layer formation on the glass surface [2]. Therefore, modifications on the SBF solution composition can be used as an indirect method to evaluate the mechanism of the HCA formation and the processes taking place on the glass surface, indicative of the glass bioactive potential [64]. Fig. 6 presents the Si, P, Ca and Co concentration after immersion of bioactive glasses in SBF for different periods of time, as evaluated by ICP-OES.

No significant difference was observed in the Si release profile between the glasses, and its concentration was stabilized in SBF solution after 24 h, maintained in the range of 62 – 70 ppm. Co was incorporated into the bioactive glass structure in partial Ca substitution, which consequently reduced the Ca content of the Co-containing samples. The concentration of Ca in SBF, however, was higher for 5Co-N samples than the BG and 5Co-C samples which presented similar behaviour. As 5Co-N presented a higher surface area than 5Co-C, it was expected that Ca release was also higher for 5Co-N. Nonetheless, the difference observed between 5Co-N and BG could be related to the formation of a weakly bonded network due to Co incorporation, which increased the dissolution rate of this sample, this is especially evident during the first 8 h of immersion [19]. The Ca release has an important role in the biomineralization process, and could also induce osteoblast proliferation, improving the biomaterial therapeutic effect [65].

The P concentration reduced immediately after sample immersion in SBF and can be related to the PO43- migration to the glass surface and the calcium phosphate-rich layer formation [66]. However, for BG, the phosphate concentration fell to near zero ppm in 24 h, whereas for Co-containing bioactive glasses the P content reduced at a lower rate, reaching a minimum of approximately 7 ppm after 2 weeks, this could be related to a slower phosphate precipitation on the glass surface, with a consequential reduction on calcium phosphate and HCA formation. Therefore, the results indicate that Co incorporation reduced the HCA layer formation rate during the two weeks of study.

Co-release showed an initial burst release, stabilizing thereafter and reaching the maximum concentration of approximately 10 and 6 ppm for 5Co-N and 5Co-C, respectively. The use of different precursors affected the ion release, as a lower content of Co was released from 5Co-C. Although the specific surface area could affect the dissolution profile of glasses, similar ion release behaviour was observed for silicon on all samples. Therefore, the difference on the Co-release observed between 5Co-C and 5Co-N during the two weeks study can be assigned to the presence of cobalt (II, III) oxide on 5Co-C sample, as it is reported that this crystalline phase is poorly soluble in aqueous medium [67].

Previous work showed that cobalt chloride at a concentration of approximately 13 ppm affected transcriptional responses to hypoxia in microvascular endothelial cells (HMEC-1), inducing the HIF-1 response [68]. Also, high VEGF expression was observed for the Co-release in the range of 3 to 12 ppm evaluated in Co-containing scaffolds of bioactive glass and collagen-glycosaminoglycan, with also increased ALP (alkaline phosphatase) activity [69]. Therefore, samples in this work presented ion release within the range of the biologically active concentration that could induce the new blood vessel formation according to previous studies. Nonetheless, recently evaluation performed in hMSC showed a reduction in chondrogenic differentiation in a time and concentration-dependent manner, although the level of HIF-1α was significantly increased [28]. Therefore, Co-release should be carefully evaluated and tailored according to the desired biomaterial application.

Fig. 6: Si, P, Ca and Co concentration in SBF solution after immersion of bioactive glass samples up to two weeks

The glass powders were evaluated by FTIR, XRD before and after the 2 weeks study, and results are shown in Fig. 7. No significant differences were observed in the FTIR spectra, in which the main bands related to the silica network can be assigned to the absorbance at around 450 cm-1 (Si-O bending), 965 cm-1 (SiO4 structures containing non-bridging oxygen) [70], and the broad band with maximum at 1020 cm-1 (Si-O stretching) [66]. The doublet at 562 and 602 cm-1 is characteristic of P-O bending of PO43- groups in apatite crystals [15], and its intensity increased after 2 weeks immersion in SBF. The band at 1060 cm-1 is often related to the stretching of P-O bonds [66]. Carbonates (CO32-) are related to the vibrational bands at 808 cm-1, 890 cm-1 [37], and the 1400 – 1470 cm-1 region [3], indicating the presence of a carbonated phosphate-rich layer on all samples. Results were further confirmed by XRD, showed in Fig. 7 (b).The appearance of pronounced peaks at 26°, 32° and 50° (2θ) can be assigned to the HCA layer formation on the glass surface after the immersion in SBF [71]. A peak attributed to calcite also was observed, and its formation can be related to the reaction between the calcium released from the glasses and the carbonate ions present in SBF solution [47]. Furthermore, the 5Co-C crystalline features remained after the dissolution study, indicating that the oxide structure was not dissolved during this study, which could explain the lower Co-release for this sample.

Fig. 7: (a) FTIR spectra and (b) XRD patterns of bioactive glasses soaked in SBF for 2 weeks.

Finally, Fig. 8 present SEM micrographs before and after 2 weeks immersion in SBF, and showed a structure with typical morphology of apatite crystal that is formed due to biomineralization on the glass surface [29,66], confirming the previous results.

Fig. 8: SEM micrographs of BG, 5Co-N and 5Co-C samples before and after 2 weeks of immersion in SBF.

Due to the dissolution results presented above, this study confirms that Co incorporation into the glass network could lead to a reduction in the HCA layer formation rate, although evidence of HCA formation was still found within 2 weeks immersion in SBF. A reduced Co-release was observed when cobalt chloride was used as a precursor, mainly related to the poorly soluble crystalline features [67]. The Co-release was within the reported to induce an angiogenic response; however, the glass composition should be properly tailored when designing new bioactive glasses for each application.

1. Conclusions

In this work, Co-releasing bioactive glasses were obtained through the sol-gel method, and the influence of Co incorporation and precursor selection on the glass structure and properties were evaluated. When cobalt chloride was used as a precursor, crystalline cobalt (II, III) oxide was formed, indicating that Co ions were not fully incorporated into the glass structure as a network modifier, whereas crystalline features were not detected when the nitrate salt was the precursor. This ion is present mainly as Co2+ species, as XPS evaluation, although Co3+ species are also evident when chloride salt was the precursor. Co-incorporation using both precursors evaluated in this work reduced the HCA layer formation rate during the two weeks immersion in SBF. A higher Co-release for the cobalt nitrate derived glass was observed, and Co release content was within the reported concentration that induces angiogenic effects, although further biological evaluation should be performed. Therefore, the Co precursor plays an important role in defining the final glass structure and properties and should be carefully evaluated when developing new sol-gel bioactive glass compositions for tissue engineering applications.

Acknowledgements

The authors gratefully acknowledge financial support from CNPq, CAPES and FAPEMIG/Brazil and the Advanced Photoelectron Spectroscopy Laboratory (APSL – Department of Materials - Imperial College London) for XPS analysis.

Conflict of interest

The authors declare that there is no personal or financial conflict of interests in the current paper.

References

[1]L.L. Hench, R.J. Splinter, W.C. Allen, T.K. Greenlee, Bonding mechanisms at the interface of ceramic prosthetic materials, J. Biomed. Mater. Res. 5 (1971) 117–141. doi:10.1002/jbm.820050611.

[2]J.R. Jones, Review of bioactive glass: From Hench to hybrids, Acta Biomater. 9 (2013) 4457–4486. doi:10.1016/j.actbio.2012.08.023.

[3]A.L.B. Maçon, S. Lee, G. Poologasundarampillai, T. Kasuga, J.R. Jones, Synthesis and dissolution behaviour of CaO/SrO-containing sol–gel-derived 58S glasses, J. Mater. Sci. 52 (2017) 8858–8870. doi:10.1007/s10853-017-0869-0.

[4]B. Yu, C. a Turdean-Ionescu, R. a Martin, R.J. Newport, J. V Hanna, M.E. Smith, J.R. Jones, Effect of Calcium Source on Structure and Properties of Sol–Gel Derived Bioactive Glasses, Langmuir. 28 (2012) 17465–17476. doi:10.1021/la303768b.

[5]V. Miguez-Pacheco, L.L. Hench, A.R. Boccaccini, Bioactive glasses beyond bone and teeth: Emerging applications in contact with soft tissues, Acta Biomater. 13 (2015) 1–15. doi:10.1016/j.actbio.2014.11.004.

[6]L.B. Romero-Sánchez, M. Marí-Beffa, P. Carrillo, M.Á. Medina, A. Díaz-Cuenca, Copper-containing mesoporous bioactive glass promotes angiogenesis in an in vivo zebrafish model, Acta Biomater. (2018). doi:10.1016/j.actbio.2017.12.032.

[7]R. Vecchione, G. Luciani, V. Calcagno, A. Jakhmola, B. Silvestri, D. Guarnieri, V. Belli, A. Costantini, P.A. Netti, Multilayered silica-biopolymer nanocapsules with a hydrophobic core and a hydrophilic tunable shell thickness, Nanoscale. 8 (2016) 8798–8809. doi:10.1039/c6nr01192f.

[8]X. Wang, W. Li, Biodegradable mesoporous bioactive glass nanospheres for drug delivery and bone tissue regeneration, Nanotechnology. 27 (2016) 1–8. doi:10.1088/0957-4484/27/22/225102.

[9]B. Silvestri, A. Pezzella, G. Luciani, A. Costantini, F. Tescione, F. Branda, Heparin conjugated silica nanoparticle synthesis, Mater. Sci. Eng. C. 32 (2012) 2037–2041. doi:10.1016/j.msec.2012.05.018.

[10]F. Tallia, L. Russo, S. Li, A.L.H. Orrin, X. Shi, S. Chen, J.A.M. Steele, S. Meille, J. Chevalier, P.D. Lee, M.M. Stevens, L. Cipolla, J.R. Jones, Bouncing and 3D printable hybrids with self-healing properties, Mater. Horizons. (2018). doi:10.1039/C8MH00027A.

[11]P. Sepulveda, J.R. Jones, L. Hench, Characterization of Melt-Derived 45S5 and sol-gel–derived 58S Bioactive Glasses, J. Biomed. Mater. Res. 58 (2001) 734–740. doi:10.1002/jbm.10026.

[12]E.G.L. Alves, R. Serakides, I.R. Rosado, M.M. Pereira, N.M. Ocarino, H.P. Oliveira, A.M. Góes, C.M.F. Rezende, Effect of the ionic product of bioglass 60s on osteoblastic activity in canines, BMC Vet. Res. 11 (2015) 247. doi:10.1186/s12917-015-0558-7.

[13]D. Arcos, M. Vallet-Regí, Sol-gel silica-based biomaterials and bone tissue regeneration., Acta Biomater. 6 (2010) 2874–88. doi:10.1016/j.actbio.2010.02.012.

[14]I.D. Xynos, A.J. Edgar, L.D.K. Buttery, L.L. Hench, J.M. Polak, Gene-expression profiling of human osteoblasts following treatment with the ionic products of Bioglass® 45S5 dissolution, J. Biomed. Mater. Res. 55 (2001) 151–157. doi:10.1002/1097-4636(200105)55:2<151::AID-JBM1001>3.0.CO;2-D.

[15]A.B. Houreh, S. Labbaf, H.-K. Ting, F. Ejeian, J.R. Jones, M.-H.N. Esfahani, Influence of calcium and phosphorus release from bioactive glasses on viability and differentiation of dental pulp stem cells, J. Mater. Sci. (2017). doi:10.1007/s10853-017-0946-4.

[16]J.R.J. Delben, O.M. Pimentel, M.B. Coelho, P.D. Candelorio, L.N. Furini, F. Alencar Dos Santos, F.S. De Vicente, A.A.S.T. Delben, Synthesis and thermal properties of nanoparticles of bioactive glasses containing silver, J. Therm. Anal. Calorim. 97 (2009) 433–436. doi:10.1007/s10973-009-0086-4.

[17] a. J. Salinas, S. Shruti, G. Malavasi, L. Menabue, M. Vallet-Regí, Substitutions of cerium, gallium and zinc in ordered mesoporous bioactive glasses, Acta Biomater. 7 (2011) 3452–3458. doi:10.1016/j.actbio.2011.05.033.

[18]M.D. O’Donnell, R.G. Hill, Influence of strontium and the importance of glass chemistry and structure when designing bioactive glasses for bone regeneration, Acta Biomater. 6 (2010) 2382–2385. doi:10.1016/j.actbio.2010.01.006.

[19]E. Gentleman, Y.C. Fredholm, G. Jell, N. Lotfibakhshaiesh, M.D. O’Donnell, R.G. Hill, M.M. Stevens, The effects of strontium-substituted bioactive glasses on osteoblasts and osteoclasts in vitro, Biomaterials. 31 (2010) 3949–3956. doi:10.1016/j.biomaterials.2010.01.121.

[20]A. Hoppe, N.S. Güldal, A.R. Boccaccini, A review of the biological response to ionic dissolution products from bioactive glasses and glass-ceramics, Biomaterials. 32 (2011) 2757–2774. doi:10.1016/j.biomaterials.2011.01.004.

[21]S. Kargozar, F. Baino, S. Hamzehlou, R.G. Hill, M. Mozafari, Bioactive Glasses: Sprouting Angiogenesis in Tissue Engineering, Trends Biotechnol. 36 (2018) 430–444. doi:10.1016/j.tibtech.2017.12.003.

[22]M. Dziadek, B. Zagrajczuk, E. Menaszek, K. Dziadek, K. Cholewa-Kowalska, A simple way of modulating in vitro angiogenic response using Cu and Co-doped bioactive glasses, Mater. Lett. 215 (2018) 87–90. doi:10.1016/j.matlet.2017.12.075.

[23]D.S. Steinbrech, B.J. Mehrara, P.B. Saadeh, J. a Greenwald, J. a Spector, G.K. Gittes, M.T. Longaker, VEGF expression in an osteoblast-like cell line is regulated by a hypoxia response mechanism., Am. J. Physiol. Cell Physiol. 278 (2000) C853–C860.

[24]M.M. Azevedo, O. Tsigkou, R. Nair, J.R. Jones, G. Jell, M.M. Stevens, Hypoxia Inducible Factor-Stabilizing Bioactive Glasses for Directing Mesenchymal Stem Cell Behavior., Tissue Eng. Part A. 00 (2014) 1–8. doi:10.1089/ten.TEA.2014.0083.

[25]M. Azevedo, G. Jell, M. O’Donnell, R. Law, R. Hill, M. Stevens, Synthesis and characterization of hypoxia-mimicking bioactive glasses for skeletal regeneration, J. Mater. Chem. 20 (2010) 8854–8864. doi:10.1039/c0jm01111h.

[26]C. Wu, Y. Zhou, W. Fan, P. Han, J. Chang, J. Yuen, M. Zhang, Y. Xiao, Hypoxia-mimicking mesoporous bioactive glass scaffolds with controllable cobalt ion release for bone tissue engineering, Biomaterials. 33 (2012) 2076–2085. doi:10.1016/j.biomaterials.2011.11.042.

[27]A. Hoppe, B. Jokic, D. Janackovic, T. Fey, P. Greil, S. Romeis, J. Schmidt, W. Peukert, J. Lao, E. Jallot, A.R. Boccaccini, Cobalt-releasing 1393 bioactive glass-derived scaffolds for bone tissue engineering applications, ACS Appl. Mater. Interfaces. 6 (2014) 2865–2877. doi:10.1021/am405354y.

[28]E. Littmann, H. Autefage, A.K. Solanki, C. Kallepitis, J.R. Jones, M. Alini, M. Peroglio, M.M. Stevens, Cobalt-containing bioactive glasses reduce human mesenchymal stem cell chondrogenic differentiation despite HIF-1α stabilisation, J. Eur. Ceram. Soc. 38 (2018) 877–886. doi:10.1016/j.jeurceramsoc.2017.08.001.

[29]R.L. Siqueira, E.D. Zanotto, The influence of phosphorus precursors on the synthesis and bioactivity of SiO2-CaO-P2O5 sol-gel glasses and glass-ceramics, J. Mater. Sci. Mater. Med. 24 (2013) 365–379. doi:10.1007/s10856-012-4797-x.

[30]H.-K. Ting, S. Page, G. Poologasundarampillai, S. Chen, J. V. Hanna, J.R. Jones, Phosphate content affects structure and bioactivity of sol-gel silicate bioactive glasses, Int. J. Appl. Glas. Sci. (2017). doi:10.1111/ijag.12322.

[31]S. Lin, C. Ionescu, K.J. Pike, M.E. Smith, J.R. Jones, Nanostructure evolution and calcium distribution in sol-gel derived bioactive glass, J. Mater. Chem. 19 (2009) 1276–1282. doi:10.1039/b814292k.

[32]A.L.B. Maçon, M. Jacquemin, S.J. Page, S. Li, S. Bertazzo, M.M. Stevens, J. V. Hanna, J.R. Jones, Lithium-silicate sol–gel bioactive glass and the effect of lithium precursor on structure–property relationships, J. Sol-Gel Sci. Technol. (2016). doi:10.1007/s10971-016-4097-x.

[33]M.M. Pereira, A.E. Clark, L.L. Hench, Homogeneity of Bioactive Sol-Gel-Derived Glasses in the System CaO-P2O5-SiO2.pdf, J. Mater. Synth. Process. 2 (1994) 189–195.

[34]R. Li, A.E. Clark, L.L. Hench, An investigation of bioactive glass powders by sol-gel processing., J. Appl. Biomater. 2 (1991) 231–239. doi:10.1002/jab.770020403.

[35]A.L.B. Maçon, T.B. Kim, E.M. Valliant, K. Goetschius, R.K. Brow, D.E. Day, A. Hoppe, A.R. Boccaccini, I.Y. Kim, C. Ohtsuki, T. Kokubo, A. Osaka, M. Vallet-Regí, D. Arcos, L. Fraile, A.J. Salinas, A. V Teixeira, Y. Vueva, R.M. Almeida, M. Miola, C. Vitale-Brovarone, E. Verné, W. Höland, J.R. Jones, A unified in vitro evaluation for apatite-forming ability of bioactive glasses and their variants, J. Mater. Sci. Mater. Med. 26 (2015) 115. doi:10.1007/s10856-015-5403-9.

[36]T. Kokubo, H. Takadama, How useful is SBF in predicting in vivo bone bioactivity?, Biomaterials. 27 (2006) 2907–2915. doi:10.1016/j.biomaterials.2006.01.017.

[37]A.A.R. de Oliveira, D.A. de Souza, L.L.S. Dias, S.M. de Carvalho, H.S. Mansur, M. de Magalhães Pereira, Synthesis, characterization and cytocompatibility of spherical bioactive glass nanoparticles for potential hard tissue engineering applications, Biomed. Mater. 8 (2013) 025011. doi:10.1088/1748-6041/8/2/025011.

[38]M.S. Hasan, U. Werner-Zwanziger, D. Boyd, Composition-structure-properties relationship of strontium borate glasses for medical applications, J. Biomed. Mater. Res. Part A. 103 (2015) 2344–2354. doi:10.1002/jbm.a.35361.

[39] a. Saboori, M. Rabiee, F. Moztarzadeh, M. Sheikhi, M. Tahriri, M. Karimi, Synthesis, characterization and in vitro bioactivity of sol-gel-derived SiO2-CaO-P2O5-MgO bioglass, Mater. Sci. Eng. C. 29 (2009) 335–340. doi:10.1016/j.msec.2008.07.004.

[40]M. Montazerian, J.F. Schneider, B.E. Yekta, V.K. Marghussian, A.M. Rodrigues, E.D. Zanotto, Sol–gel synthesis, structure, sintering and properties of bioactive and inert nano-apatite–zirconia glass–ceramics, Ceram. Int. 41 (2015) 11024–11045. doi:10.1016/j.ceramint.2015.05.047.

[41]S. Yuvaraj, L. Fan-yuan, C. Tsong-huei, Y. Chuin-tih, Thermal Decomposition of Metal Nitrates in Air and Hydrogen Environments, (2003) 1044–1047. doi:10.1021/jp026961c.

[42]J.R. Jones, L.M. Ehrenfried, L.L. Hench, Optimising bioactive glass scaffolds for bone tissue engineering, Biomaterials. 27 (2006) 964–973. doi:10.1016/j.biomaterials.2005.07.017.

[43]H. Pirayesh, J. a. Nychka, Sol-gel synthesis of bioactive glass-ceramic 45S5 and its in vitro dissolution and mineralization behavior, J. Am. Ceram. Soc. 96 (2013) 1643–1650. doi:10.1111/jace.12190.

[44]A.M. El-Kady, A.F. Ali, Fabrication and characterization of ZnO modified bioactive glass nanoparticles, Ceram. Int. 38 (2012) 1195–1204. doi:10.1016/j.ceramint.2011.07.069.

[45]C.D.F. Moreira, S.M. Carvalho, H.S. Mansur, M.M. Pereira, Thermogelling chitosan–collagen–bioactive glass nanoparticle hybrids as potential injectable systems for tissue engineering, Mater. Sci. Eng. C. 58 (2016) 1207–1216. doi:10.1016/j.msec.2015.09.075.

[46]M. Catauro, A. Dell, S. Vecchio, Synthesis , structural , spectroscopic and thermoanalytical study of sol – gel derived SiO 2 – CaO – P 2 O 5 gel and ceramic materials, Thermochim. Acta. 625 (2016) 20–27. doi:10.1016/j.tca.2015.12.004.

[47]I. Atkinson, E.M. Anghel, L. Predoana, O.C. Mocioiu, L. Jecu, I. Raut, C. Munteanu, D. Culita, M. Zaharescu, Influence of ZnO addition on the structural, in vitro behavior and antimicrobial activity of sol–gel derived CaO–P2O5–SiO2 bioactive glasses, Ceram. Int. 42 (2016) 3033–3045. doi:10.1016/j.ceramint.2015.10.090.

[48]J. Serra, P. González, S. Liste, C. Serra, S. Chiussi, B. León, M. Pérez-Amor, H.O. Ylänen, M. Hupa, FTIR and XPS studies of bioactive silica based glasses, J. Non. Cryst. Solids. 332 (2003) 20–27. doi:10.1016/j.jnoncrysol.2003.09.013.

[49]K.J. Rao, H. Benqlilou-Moudden, B. Desbat, P. Vinatier, A. Levasseur, Infrared spectroscopic study of LiCoO2thin films, J. Solid State Chem. 165 (2002) 42–47. doi:10.1006/jssc.2001.9487.

[50]D.C.L. Vasconcelos, E.H.M. Nunes, M. Houmard, J. Motuzas, J.F. Nascimento, W. Grava, V.S.T. Ciminelli, J.C. Diniz, W.L. Vasconcelos, Structural investigation of cobalt-doped silica derived from sol – gel synthesis, J. Non. Cryst. Solids. 378 (2013) 1–6. doi:10.1016/j.jnoncrysol.2013.06.008.

[51]G. Ortega-Zarzosa, C. Araujo-Andrade, M.E. Compeán-Jasso, J.R. Martínez, F. Ruiz, Cobalt Oxide / Silica Xerogels Powders : X-Ray Diffraction , Infrared and Visible Absorption Studies, J. Sol-Gel Sci. Technol. 24 (2002) 23–29. doi:10.1023/A:101510541.

[52]V. Musat, E. Fortunato, A.M. Botelho, R. Monteiro, Sol – gel cobalt oxide – silica nanocomposite thin films for gas sensing applications, 516 (2008) 1499–1502. doi:10.1016/j.tsf.2007.07.197.

[53]L.O. Simonsen, H. Harbak, P. Bennekou, Cobalt metabolism and toxicology-A brief update, Sci. Total Environ. 432 (2012) 210–215. doi:10.1016/j.scitotenv.2012.06.009.

[54]A.M. Toufiq, F. Wang, Q. Javed, Y. Li, Influence of SiO2 on the structure-controlled synthesis and magnetic properties of prismatic MnO2 nanorods., Nanotechnology. 24 (2013) 415703. doi:10.1088/0957-4484/24/41/415703.

[55]J. Dupin, D. Gonbeau, P. Vinatier, A. Levasseur, Systematic XPS Studies of Metal Oxides , Hydroxides and Peroxides Systematic XPS studies of metal oxides , hydroxides and peroxides, Phys. Chem. Chem. Phys. 2 (2000) 1319–1324. doi:10.1039/a908800h.

[56]J. Perez-Pariente, F. Balas, M. Vallet-Regi, Surface and chemical study of SiO2.P2O5.CaO.(MgO) bioactive glasses, Chem. Mater. 12 (2000) 750–755. doi:10.1021/cm9911114.

[57]A. Mekki, M. Salim, XPS study of transition metal doped silicate glasses, J. Electron Spectros. Relat. Phenomena. 101 (1999) 227–232. doi:10.1016/S0368-2048(98)00450-2.

[58]M.C. Biesinger, B.P. Payne, A.P. Grosvenor, L.W.M. Lau, A.R. Gerson, R.S.C. Smart, Resolving surface chemical states in XPS analysis of first row transition metals, oxides and hydroxides: Cr, Mn, Fe, Co and Ni, Appl. Surf. Sci. 257 (2011) 2717–2730. doi:10.1016/j.apsusc.2010.10.051.

[59]K. Elkabouss, M. Kacimi, M. Ziyad, S. Ammar, F. Bozon-Verduraz, Cobalt-exchanged hydroxyapatite catalysts: Magnetic studies, spectroscopic investigations, performance in 2-butanol and ethane oxidative dehydrogenations, J. Catal. 226 (2004) 16–24. doi:10.1016/j.jcat.2004.05.007.

[60]T.J. Chuang, C.R. Brundle, D.W. Rice, Interpretation of X-Ray Photoemission Spectra of Cobalt Oxides and Cobalt Oxide Surfaces, Surf. Sci. 59 (1976) 413–429.

[61]A. Mekki, D. Holland, K. Ziq, C.F. Mcconville, XPS and magnetization studies of cobalt sodium silicate glasses, J. Non. Cryst. Solids. 3 (1997) 267–279.

[62]E. Kramer, E. Itzkowitz, M. Wei, Synthesis and characterization of cobalt-substituted hydroxyapatite powders, Ceram. Int. 40 (2014) 13471–13480. doi:10.1016/j.ceramint.2014.05.072.

[63]M.M. Pereira, A.E. Clark, L.L. Hench, Effect of Texture on the Rate of Hydroxyapatite Formation on Gel-Silica Surface, J. Am. Ceram. Soc. 78 (1995) 2463–68.

[64]A.A.R. de Oliveira, B.B. de Carvalho, H. Sander Mansur, M. de Magalhães Pereira, Synthesis and characterization of bioactive glass particles using an ultrasound-assisted sol–gel process: Engineering the morphology and size of sonogels via a poly(ethylene glycol) dispersing agent, Mater. Lett. 133 (2014) 44–48. doi:10.1016/j.matlet.2014.06.092.

[65]S. Murphy, D. Boyd, S. Moane, M. Bennett, The effect of composition on ion release from Ca-Sr-Na-Zn-Si glass bone grafts, J. Mater. Sci. Mater. Med. 20 (2009) 2207–2214. doi:10.1007/s10856-009-3789-y.

[66]J. Bejarano, P. Caviedes, H. Palza, Sol–gel synthesis and in vitro bioactivity of copper and zinc-doped silicate bioactive glasses and glass-ceramics, Biomed. Mater. 10 (2015) 025001. doi:10.1088/1748-6041/10/2/025001.

[67]C. Uboldi, T. Orsière, C. Darolles, V. Aloin, V. Tassistro, I. George, V. Malard, Poorly soluble cobalt oxide particles trigger genotoxicity via multiple pathways, Part. Fibre Toxicol. 13 (2015) 5. doi:10.1186/s12989-016-0118-8.

[68] a Minchenko, J. Caro, Regulation of endothelin-1 gene expression in human microvascular endothelial cells by hypoxia and cobalt: role of hypoxia responsive element., Mol. Cell. Biochem. 208 (2000) 53–62.

[69]E. Quinlan, S. Partap, M.M. Azevedo, G. Jell, M.M. Stevens, F.J. O’Brien, Hypoxia-mimicking bioactive glass/collagen glycosaminoglycan composite scaffolds to enhance angiogenesis and bone repair, Biomaterials. 52 (2015) 358–366. doi:10.1016/j.biomaterials.2015.02.006.

[70]C. Paluszkiewicz, A. Ślósarczyk, D. Pijocha, M. Sitarz, M. Bućko, A. Zima, A. Chróścicka, M. Lewandowska-Szumieł, Synthesis, structural properties and thermal stability of Mn-doped hydroxyapatite, J. Mol. Struct. 976 (2010) 301–309. doi:10.1016/j.molstruc.2010.04.001.

[71]Y.C. Fredholm, N. Karpukhina, D.S. Brauer, J.R. Jones, R. V Law, G. Hill, Y.C. Fredholm, N. Karpukhina, D.S. Brauer, Influence of strontium for calcium substitution in bioactive glasses on degradation , ion release and apatite formation Influence of strontium for calcium substitution in bioactive glasses on degradation , ion release and apatite formation, (2011).

Highlights

· The Co precursor selection can affect the bioactive glass properties

· Crystalline structures can be found when cobalt chloride is used as Co precursor

· Hydroxycarbonate apatite can be formed on Co-containing bioactive glass surface

Keywords

bioactive glass; sol-gel; cobalt; bioactivity

Figure Captions

Fig. 1: (a) Thermogravimetric analysis (TGA) and (b) differential scanning calorimetry (DSC) of bioactive glasses.

Fig. 2: (a) Fourier transform infrared (FTIR) spectra and (b) X-Ray Diffraction (XRD) patterns of bioactive glass samples.

Fig. 3: XPS spectra of bioactive glass samples.

Fig. 4: High-resolution XPS spectra of (a) Si 2p (b) O 1s and (c) Co 2p for bioactive glasses samples

Fig. 5: (a) Nitrogen sorption isotherms and (b) pore size distribution (BJH algorithm on desorption branches) of bioactive glass samples.

Fig. 6: Si, P, Ca and Co concentration in SBF solution after immersion of bioactive glass samples up to two weeks

Fig. 7: (a) FTIR spectra and (b) XRD patterns of bioactive glasses soaked in SBF for 2 weeks.

Fig. 8: SEM micrographs of BG, 5Co-N and 5Co-C samples before and after 2 weeks of immersion in SBF.

Table Captions

Table 1: Nominal and measured composition of bioactive glasses (% mol), determined by lithium metaborate fusion and ICP-OES.

Table 2: Summary of the textural characterization obtained performed on bioactive glasses

1