Cold hydrostatic sintering of 45S5 bioactive glass

Abstract Fully amorphous 45S5 bioactive glass pellets were successfully consolidated at room temperature using cold hydrostatic sintering (CHS) and their acellular bioactivity was determined by immersion test in simulated body fluid (SBF). Crystalline 45S5 BG samples prepared using conventional sintering (TS) at 1050 °C were also tested. Preliminary results indicated the formation of hydroxyapatite (HAP) on the surfaces of both samples after immersion in SBF, which was confirmed using SEM, FTIR, and XRD analyses. The amount of hydroxyapatite increases by increasing the SBF immersion time. Results revealed that the amount of HAP crystalline phase was similar in both samples after 14 days of incubation in SBF (i.e., 23% in TS and 25% in CHS samples). The pellets prepared by TS resulted in a relative density of 95.6% and a hardness of about 3.5 GPa. On the other hand, CHS samples had a hardness of 1.2 GPa and a relative density of 75.1%, indicating the presence of a porous structure that can be beneficial for the interaction of the material with proteins and cells (both in vitro and in vivo).


Introduction
45S5 bioactive glass (BG) is a well-established material for bone tissue engineering (BTE) applications due to its strong bonding ability with host bone tissue in a physiological environment, which results from the formation of bone-like hydroxycarbonate apatite (HCA) layer on the BG surface (Baino, Hamzehlou, & Kargozar, 2018;Blaeß, Müller, Poologasundarampillai, & Brauer, 2019;Hench, 2006).The development of BGs with enhanced chemical reactivity under physiological conditions (in vitro and in vivo), is important in the context of BTE applications (Grasso, Chinnam, Porwal, Boccaccini, & Reece, 2013), in particular considering the tendency to crystallize of 45S5 BG, which could reduce the kinetics of HCA formation (Chen, Thompson, & Boccaccini, 2006).
It has been reported that amorphous 45S5 BG exhibits higher dissolution rate and increased ion release than crystallized BG (Montinaro et al., 2018).The sintering of 45S5 BGs above the glass transition temperature (about 600 °C) leads to crystallization.It has been reported that both the rate of glass dissolution and the ion release kinetics in crystallized 45S5 BG are lower than those of amorphous 45S5 BG (Bretcanu et al., 2009;Fiume, Barberi, Verné, & Baino, 2018;Hoppe, Guldal, Boccaccini, Güldal, & Boccaccini, 2011).In addition, it is known that crystallization retards bioactivity (Chen et al., 2006;Jones, 2013).Moreover, Filho, Latorre, and Hench (1996) found that the HCA formation onset time of amorphous BG is almost three times lower (i.e., 8 h) than the one of crystallized BG (20 h).On the other hand, it is not possible to obtain a fully amorphous and dense 45S5 BG structure using a conventional sintering approach.
Apart from preventing the crystallization of the BG, room temperature consolidation might enable the effective incorporation of organic additives (i.e., polymers) and therapeutic compounds that would be decomposed when heated at high temperatures in a conventional (high-temperature) sintering process.Low-temperature densification processes, such as the cold sintering process (CSP), allow the densification of various inorganic powders at room temperature exploiting the chemical affinity between solute and solvent (Taveri et al., 2021;Vakifahmetoglu & Karacasulu, 2020).For this reason, this paper investigates the relationship between the cold sintering process of 45S5 BG and the resulting in vitro bioactivity in simulated body fluid (SBF) as a starting study to assess the convenience of the cold hydrostatic sintering approach for 45S5 BG. 45S5 BG pellets were prepared by cold hydrostatic sintering under an isostatic pressure in the presence of water.Sintered 45S5 BG pellets were also prepared using uniaxial pressing followed by conventional sintering at 1050 °C.

Experimental procedure
45S5 BG powder (45 SiO 2 , 24.5 Na 2 O, 24.5 CaO, and 6 P 2 O 5 , wt.%) of ≤26 µm size was purchased from Chengdu Dikang Zhongke Biomedical Material Co., Ltd, China.BG powder and water were loaded into a sealed vial, the mixture was homogenized using a vortex mixer (XH-C, Lianer Corp., China) at 2000 rpm for 10-20 min.The selected volume ratio of water was 20% as it resulted in a relative density of ≈75 ± 3%.

Cold sintering experiment
The cold hydrostatic sintering approach was initially proposed by Jiang at al. (Jiang et al., 2019).First, the mixture was placed into a steel mould whose inner wall was coated with a lubricating agent.This step was to obtain a specimen with a regular cylindrical shape.Subsequently, the discs were pressed using an isostatic press (Shanxi Technology Co., Ltd) under 300 MPa for 60 min.The samples were slowly dried at 40 °C in a vacuum (0.1 mbar) oven for 7 days.This was done to prevent carbonation resulting from the reaction with atmospheric CO 2 .For comparison, samples were sintered in a traditional muffle furnace at 1050 °C with a heating rate of 5 °C/min, and a holding time of 2 h followed by natural cooling.All processing parameters are summarized in Table 1.

Characterization of the pellets
The density of the samples was determined using an electron densitometer (ZMD-2 Series electronic density meter, NanTong Precision Instrument Co., Ltd.).The hardness was measured using a Vickers hardness tester (HVS-50 Digital Vickers Hardness Tester, Shanghai Wanheng Precision Instrument Co., Ltd.) with an applied load of 1 kg and a dwell time of 15 s.The samples were immersed in water for several days to verify their stability.The Vickers hardness was measured before and after immersion to quantify the possible decay due to water absorption.
The in-vitro mineralization behavior of samples was investigated by immersing the samples in simulated body fluid (SBF) as proposed by Kokubu and Takadama (2006).The pellets were immersed in SBF (pH 7.4) for set time intervals (1 h, 6 h, 7 days, and 14 days).Samples were then removed, rinsed with mili-Q water and dried.Afterward, samples were characterized using FE-SEM (FESEM, LEO 435VP, Carl Zeiss™ AG) and XRD (D8 Advance, Brucker AXS, Karlsruhe, Germany) analysis to characterize the morphological and structural changes on the surface of the pellets.After XRD measurements, quantitative analysis of the crystalline phases was carried out by Rietveld refinement with the software TOPAS V5 (Bruker AXS, Karlsruhe, Germany) and subsequent application of the G-factor, a standard external method.

Density and hardness measurements
The density of the samples prepared by the different processes is presented in Table 2.The theoretical density of the 45S5 BG powder used for the preparation of pellets is 2.7 g/cm 3 .The bioactive glass prepared by traditional sintering had a relative density exceeding 95%.The pellets sintered under cold isostatic pressing exhibited a lower density of 75%.
As suggested by Taveri et al. higher densities are possible by adjusting the ratio between water and solid fractions (Taveri et al., 2021).It is worth noticing that using conventional sintering, a density of ∼75% in 45S5 BG can be reached only at T > 800 °C (Blaeß et al., 2019).
The hardness of both prepared pellets was measured by the Vickers hardness method.The pellets prepared by using traditional hightemperature sintering have the highest hardness, about 3.5 GPa.The hardness is decreased to 1.16 GPa in the samples prepared using CHS.This reduction can be ascribed to the ≈25% residual porosity.The obtained hardness value of CHS samples is higher than the hardness of human bone (0.6 GPa).In addition to that, the water content used during the cold sintering process is harmless to the human body, breaking the shortcomings of potential low performance of crystallized BG, which further increases the scope of CHS-processed BGs for bone tissue engineering applications.
Furthermore, the stability of the pellets was also examined by immersing them in water for 7 days.It was observed that pellets were not dissolved after soaking in water for seven days, whereas the hardness dropped from 1.16 to 0.76 GPa.

Bioactivity evaluation
The ability of the samples to form a strong bond with bone is usually evaluated by determining sample bioactivity, i.e., the potential to develop a biologically active HCA layer on the sample surface upon immersion in simulated body fluid.The HCA formation mechanism in bioactive glasses and glass ceramics is well-established and reported in the literature (Baino et al., 2018).The present study thus aimed to investigate a new processing route for manufacturing amorphous 45S5 BG pellets that should promote HCA formation, avoiding high-temperature sintering which leads to BG crystallization (Bretcanu et al., 2009).
Figure 1 shows SEM images of the surface of the pellets before and after immersion in SBF.SEM images were taken at two different magnifications to assist the comparisons.Dissolution and surface changes in all samples upon immersion in SBF were observed.More specifically, apatite-like crystals were formed on the surface of TS and CHS samples.FTIR was used to investigate the growth of HCA layers on the BG pellet surfaces.Figure 2 displays the FITR spectra of BG-CHS and BG-TS samples after incubation in SBF for up to 14 days.In samples immersed in SBF, new transmittance double peaks are observed at 560 cm −1 , 600 cm −1 and 1014 cm −1 (Balasubramanian et al., 2016;Grasso et al., 2013), which are assigned to P-O vibration (phosphate groups), and a small peak at 870 cm −1 (Balasubramanian et al., 2016), which is associated with the carbonate band.The transmittance band of P-O group has risen with immersion time, showing that more HA has appeared on the surface of the BG-CHS and BG-TS samples.According to these findings, both samples showed types in vitro bioactivity in SBF.
Moreover, XRD analysis was carried out to investigate the crystalline nature of the newly developed phase.The formed crystalline phases were quantified using Rietveld refinement, and their crystalline fractions are shown in Table 3. Figure 3 shows the XRD patterns of all samples after immersion in SBF, and the relevant crystalline phases are also marked.
In TS sample, reflections of combeite (Na 2 Ca 2 Si 3 O 9 ) and hydroxyapatite were identified after 1 h and 6 h.It can be seen from the XRD patterns that the fraction of HAP phase increases with an increase in SBF soaking time up to 7d.In contrast to this, the fraction of combeite decreases with the SBF immersion time.This could be due to the surface dissolution and ion exchange which resulted in the formation of HAP phase as discussed elsewhere (Chen et al., 2006).The strong preferred orientation of HAP crystals in [112] direction was also noticed.
In contrast, no crystalline reflections were visible after 1 h in CHS samples.However, after 6 h, minor reflections were observed that matched with HAP and brucite (Mg(OH) 2 ) phases.Some traces of Mg(OH) 2 could   be due to the precipitation of Mg ions from the SBF solution or impurity/contaminations.With increasing SBF immersion time, HAP crystals were grown as depicted by the increase in the intensity of their relevant XRD peaks.A strong preferred orientation of HAP in [010] direction was confirmed.Furthermore, the formed crystalline phases were quantified using Rietveld refinement and the quantification of the crystalline phases is depicted in Table 3.In CHS samples, HA crystals were observed to form after 6 h of incubation.HAP crystals grew with increasing immersion time and 25 wt.% of HAP crystals was quantified until 14 days (approximated using G-factor as mentioned in (Nawaz et al., 2020)).In TS samples, an indication of HAP formation was noticed after 1 h of SBF immersion, this phase continued to grow with immersion time, and 23 wt.% of HAP crystals was detected after 14 days.Dissolution of the combeite and Na 3 H(SO 4 ) 2 phases occurred.
In summary, the bioactivity results of amorphous (CHS) and crystallized (TS) BG samples suggest that initially (SBF immersion time up to 7 days), HAP formation was more pronounced in TS samples.However, after 2 weeks of SBF immersion, the amount of HAP crystalline phase was almost similar in both samples, i.e., 23% in TS and 25% in CHS samples.A minor difference in HAP crystalline phase could be due to the higher dissolution of the CHS sample, indicating a higher bioreactivity of amorphous BG compared to crystallized BG in SBF.Similar results were found by Plewinski et al. (2013), suggesting that dissolution rate and bioactivity can be adjusted by controlling the BG crystallinity.A direct comparison of the samples is here not possible, however, due to the different densities of pellets.

Conclusions
By using a cold hydrostatic sintering approach, it was possible to obtain relatively (75% density) dense and fully amorphous 45S5 BG pellets, which might be exploited to produce bioactive glass matrix composites avoiding the need for a high-temperature densification step.The results showed that crystalline HAP was formed on both TS and CHS samples after immersion in simulated body fluid under physiological conditions.The amount of HAP crystalline phase was almost similar in both samples after 14 days of incubation (i.e., 23 wt.% in TS and 25 wt.% in CHS samples).The in vitro (SBF) dissolution study revealed that the dissolution rate and bioactivity can be adjusted by controlling sample crystallinity.Overall, CHS appears as a convenient technique to consolidate BG powder at room temperature, leading to mechanically robust samples (hardness = 1.16 GPa) and high acellular in vitro bioactivity.

Figure 1 .
Figure 1.sEM images of CHs and ts samples after immersion in sBf, showing a change in surface morphology with immersion time.

Figure 2 .
Figure 2. ftir patterns of the ts and CsH samples after immersion in sBf for upto 14 days.relevant peaks are mentioned and discussed in the text.a) BG-ts and b) BG-CHs samples.

Figure 3 .
Figure 3. XrD patterns of the ts and CsH samples after immersion in sBf for up to 14 days, showing the formation of crystalline phases with immersion time.a) BG-CHs and b) BG-ts samples.

Table 1 .
Detail of experiments and sample labelling.

Table 2 .
Density of the samples prepared by different processes.

Table 3 .
the quantification of formed crystalline phases obtained by rietveld refinement.the mass attenuation coefficient of the samples was calculated with the chemical composition of 45s5 BG, excluding the precipitated Hap. *