Ceramic Stereolithography of Bioactive Glasses: Influence of Resin Composition on Curing Behavior and Green Body Properties

Herein we report on the preparation of a bioactive glass (BAG)-based photocurable resin for the additive manufacturing of BAG scaffolds with high filler loadings. The preparation of glass/ceramics resins for stereolithography with high filler loading is always a challenge, especially for fillers with a high refractive index variance. Various photocurable resin compositions with and without bioactive glass fillers have been investigated to see the influence of bioactive glass on physical properties of the resin and resulting green body. The effect of concentration of monomers, reactive diluent, light absorber (Sudan orange G dye), photoinitiator (PI), non-reactive diluent, and fillers (BAG) on rheology and photocuring behavior of the resin and tomography of the resulting 3D structures have been investigated. The BAG contents affect the rheology of resin and influence the rate of the polymerization reaction. The resin compositions with 55–60% BAG, 10% PEG-200 (diluent), 1% of PI and 0.015% of the dye were found to be suitable compositions for the stereolithographic fabrication. A higher percentage of PI caused over-curing, while a higher amount of dye decreased the cure depth of the resin. The micro-computed tomography (µ-CT) and scanning electron microscopic (SEM) images of the resulting green bodies display a relatively dense glass scaffold without any visible cracks and good interlayer connection and surface finishing. These properties play an important role in the mechanical behavior of 3D scaffolds. This study will be helpful to prepare high density glass/ceramic slurries and optimize their printing properties.


Introduction
Bioactive glass (BAG) is a biomaterial with the ability to form a chemical bond with bone through the formation of carbonated hydroxyapatite [1]. It has been extensively used as bone and dental filler in biomedical applications due to its excellent biocompatibility, bioactivity, and osteoconductivity [2]. The use of BAG at load-bearing sites is

Materials
Bioactive glass powder with a mean diameter of 4 µm (d50 = 4 ± 1 µm) and a specific surface area of 1.5 m 2 /g was purchased from Schott (Mainz, Germany). Trimethylolpropane ethoxylate triacrylate (TMPE-OTA CQ is a type II photoinitiator which functions by intramolecular abstraction from tertiary amines [18]. It absorbs light at a wavelength longer than 400 nm and is applicable for the visible light region. A combination of CQ with co-initiator DMAB has been reported to be biocompatible and is used in many dental resins [19]. TMPE-OTA was used as a base monomer due to its low viscosity and high reactivity, while HEA was employed as a reactive diluent. The physical properties of the materials applied in this work are given in Table S1.

Preparation of Photocurable Resins
The photocurable resin was prepared using monomers TMPE-OTA and HEA in various ratios, CQ (photo-initiator) and DMAB (co-initiator) with a CQ:DMAB weight ratio = 1:1, Sudan Orange G dye (as a light absorber) and PEG-200 (as a non-reactive diluent and rheology modifier). Various loadings of BAG filler (20-65 wt.%) were added into the resin. The final mixture was homogenized in a milling vessel using a roller bank at a speed of 50 min −1 for 18 h of milling. An overview of different resin compositions investigated in this study is given in Tables 1 and 2. The resin composition was described as B a M b PI c D d E e , where the letters B, M, PI, D and E represent the BAG, monomers, photo-initiator, dye and diluent (PEG-200), respectively. The small letters in subscript, i.e., a, b, c, d and e show the concentration/ weight fraction of these components as explained below:  Tables 1 and 2, and Tables S2 and S3 in supplementary information.

Fabrication of BAG Scaffolds
A Lithoz CeraFab 7500 (Lithoz GmbH, 1060 Vienna, Austria) system was used for the photopolymerization. The wavelength range of the light source was between 400 and 500 nm. The thickness of the layer of the printed scaffold was set to 25 µm and lateral resolution reached 40 µm. The light exposure time of the first five layers was 3900 ms and for the following layers it was 3500 ms.
However, the initial curing reactions for optimization of the concentration of monomers, PI, dye and PEG-200 were carried out using an in-house developed set-up shown in Figure 1. The setup consisted of a hollow metallic cylinder with an inner diameter of 72 mm and with an LED (3 W, 460-470 nm) located on the top of the cylinder. The light exposure time was varied between 10-180 s. The cure depth was measured by a Vernier caliper (n = 5) after washing the cured samples with isopropanol to remove the uncured monomers [20].
However, the initial curing reactions for optimization of the concentration of monomers, PI, dye and PEG-200 were carried out using an in-house developed set-up shown in Figure 1. The setup consisted of a hollow metallic cylinder with an inner diameter of 72 mm and with an LED (3 W, 460-470 nm) located on the top of the cylinder. The light exposure time was varied between 10-180 s. The cure depth was measured by a Vernier caliper (n = 5) after washing the cured samples with isopropanol to remove the uncured monomers [20].

Characterization Techniques
The particle size distribution of the BAG was analyzed on a Beckman Coulter LS 13 320 laser diffraction particle size analyzer (Brea, MA, USA) in Universal Liquid Mode (ULM). The powders were dispersed in water using a probe sonicator to obtain a homogeneous distribution of particles in the water and avoid agglomeration. The particle size measurement was performed in terms of volume percentage (vol.%).
The rheology was determined by using a Physica MCR 301 Rheometer (Anton Paar, Graz, Austria). The PP 25 parallel plate geometry (Ø = 25 mm) was used to measure the rheological behavior of each polymer and resin compositions. The gap between resin and plate was set to 0.5 mm, with the shearing rate increasing linearly from 1 to 200 s −1 or 600 s −1 at a ramp rate of 1 s −1 .
Photo-Differential Scanning Calorimetry (DSC) under exposure to UV light in the range between 250 and 600 nm at 10, 50 and 100 mW/s was conducted to measure heat flow during curing. The measurements were performed by using a DSC Q2000 (TA Instruments, Milford, MA, USA) calorimeter.
Imaging of the green body was performed by Scanning Electron Microscopy (SEM) employing a Gemini LEO 1530 FEG-SEM (Carl Zeiss AG, Oberkochen, Germany) with a Noran EDX system by Thermo Fisher Scientific.
Structural analysis was carried out using micro-computed tomography (µCT) scanning on a Phoenix Nanotom M (General Electrics, Boston, MA, USA) at a voxel size of 14.99 µm. The scan was carried out at a voltage of 140 kV, a current of 160 µA, using large tube mode (0) and without the usage of filters.

Selection of the Acrylate Composition and Dye Content in the Resin
Photocurable resins comprise reactive monomers (in our experiment two different acrylates monomers TMPE-OTA and HEA), PI (CQ-photoinitiator and DMAB-co-initiator) and a light absorber (Sudan orange G dye). To understand the role of each component in the resin different resin, the compositions M b PI c D d without bioactive glass filler were prepared and investigated.
In the first step, the influence of the acrylate composition (meaning ratio of TMPE-OTA and HEA monomers) on resin viscosity and curing depth of the resin was investigated ( Figure 2). It has been observed that with increasing HEA content in M b PI c (without dye and bioactive glass) samples, the viscosity of the resin (Figure 2, viscosity measured at shear rate 200 s −1 ) rapidly decreased. This is due to HEA being a smaller monomer with a lower density and initial viscosity compared to TMPE-OTA (Table S1). The lowest viscosity observed was 10 mPa.s for M 30 PI 1 , while for M 90 PI 1 with the highest HEA content, the viscosity was 56 mPa.s. The monomer contents also had a significant influence on the photocuring behavior of the specimens. Figure 2 also shows that the value of cure depth increases with an increase in the HEA content in resin up to M 70 PI 1 (30% HEA), while a further increase led to gel-like behavior for both M 30 PI 1 and M 50 PI 1 due to a higher content of less viscous HEA and a low percentage of the triacrylate TMPE-OTA. The resins M90PI1, M80PI1 and M70PI1 were selected for further investigations, because other compositions showed semi-solid/gel-like behavior during the curing as discussed above. In situ UV-DSC was used to understand the effect of light intensity and HEA content on photopolymerization of acrylates. Three different intensities of UV light, i.e., 10, 50 and 100 mW/cm 2 ( Figure 3A), were used for in situ photo-curing of the sample M80PI1D0.01, to determine the effect of light intensity as a function of time on photopolymerization of acrylates. The amount of exothermic heat tends to increase as the light intensity increases as shown in Figure 3A [21]. The time that the resin reaches the exothermic peak represents the earlier starting point. Figure 3A shows the high reaction speed at 10 mW/cm 2 intensity reaches maxima at 1.6 min, while at 50 mW/cm 2 , it took 1.0 min. The reaction induction time reduces during the curing at higher light intensity depicting stronger exothermic reaction, which may entrap many unreacted radicals. Therefore, for further investigations, 50 mW/cm 2 light intensity was used for the polymerization reactions. Figure 3B shows the time-dependent heat release profiles for several resin compositions at 50 mW/cm². The amount of heat released, as determined by the area beneath the peak, for samples M70PI1, M80PI1, and M90PI1 (with increasing HEA contents), were 357.4, 322.3, and 313.2 J/g, respectively. Thus, the amount of energy released was almost proportional to the HEA contents. In contrast, the rate of polymerization decreased with an increase of HEA ratio. As the TMPE-OTA molecule possesses three acrylate groups, a higher TMPE-OTA content resulted in accelerated photopolymerization and higher heat flow in M90PI1 (15 W/g, 0.75 min) as compared to M70PI1 (7 W/g, 1.25 min) [22]. The resins M 90 PI 1 , M 80 PI 1 and M 70 PI 1 were selected for further investigations, because other compositions showed semi-solid/gel-like behavior during the curing as discussed above. In situ UV-DSC was used to understand the effect of light intensity and HEA content on photopolymerization of acrylates. Three different intensities of UV light, i.e., 10, 50 and 100 mW/cm 2 ( Figure 3A), were used for in situ photo-curing of the sample M 80 PI 1 D 0.01 , to determine the effect of light intensity as a function of time on photopolymerization of acrylates. The amount of exothermic heat tends to increase as the light intensity increases as shown in Figure 3A [21]. The time that the resin reaches the exothermic peak represents the earlier starting point. Figure 3A shows the high reaction speed at 10 mW/cm 2 intensity reaches maxima at 1.6 min, while at 50 mW/cm 2 , it took 1.0 min. The reaction induction time reduces during the curing at higher light intensity depicting stronger exothermic reaction, which may entrap many unreacted radicals. Therefore, for further investigations, 50 mW/cm 2 light intensity was used for the polymerization reactions. Figure 3B shows the time-dependent heat release profiles for several resin compositions at 50 mW/cm 2 . The amount of heat released, as determined by the area beneath the peak, for samples M 70 PI 1 , M 80 PI 1 , and M 90 PI 1 (with increasing HEA contents), were 357.4, 322.3, and 313.2 J/g, respectively. Thus, the amount of energy released was almost proportional to the HEA contents. In contrast, the rate of polymerization decreased with an increase of HEA ratio. As the TMPE-OTA molecule possesses three acrylate groups, a higher TMPE-OTA content resulted in accelerated photopolymerization and higher heat flow in M 90 PI 1 (15 W/g, 0.75 min) as compared to M 70 PI 1 (7 W/g, 1.25 min) [22]. Biomedicines 2022, 10, x FOR PEER REVIEW 7 of 18 Dye in photocurable resins is used to absorb photons (without creating free radicals) and suppress the light scattering that may occur due to deflection by filler particles over the slurry layers, which helps to control the overgrowth [23]. Therefore, the influence of the dye content on the photocuring behavior of the resin was investigated for a more reactive acrylate composition M90PI1. Figure 4 shows that the dye content has an influence on the cure depth of the resin at 10 and 30 s exposure time. In M90PI1D0, curing took place after 10 s exposure, while both M90PI1D0.005 and M90PI1D0.01 remained in gel form at this exposure time. On increasing the exposure time to 30 s, the cure depth of M90PI1D0, M90PI1D0.005 and M90PI1D0.01 was 2.02, 1.45 and 0.45 mm, respectively. Hence, samples with more dye content showed less curing depth. When the exposure time increases to 60 s, curing volume shrinkage was observed in the M90PI1D0 (sample without dye). For both M90PI1D0.005 and M90PI1D0.01 samples with different dye contents, the curing depth value increases at 60 s exposure time because the available free radicals which may cause volume shrinkage are readily absorbed by dye. Dye in photocurable resins is used to absorb photons (without creating free radicals) and suppress the light scattering that may occur due to deflection by filler particles over the slurry layers, which helps to control the overgrowth [23]. Therefore, the influence of the dye content on the photocuring behavior of the resin was investigated for a more reactive acrylate composition M 90 PI 1 . Figure 4 shows that the dye content has an influence on the cure depth of the resin at 10   Based on above mentioned results, in the first instance-based on the higher photopolymerization rates ( Figure 3B)-the compositions M80PI1 and M90PI1 with a higher TMPE-OTA content were selected for the formulation of resins with BAG particles. In addition, though the dye content showed no direct effect on photopolymerization of the monomers, it clearly reduced the overcuring. Therefore, the quantity of dye in the following experiments was related to the mass fraction of BAG, i.e., the higher the percentage of BAG, the higher the amount of dye required to control the light scattering. Based on above mentioned results, in the first instance-based on the higher photopolymerization rates ( Figure 3B)-the compositions M 80 PI 1 and M 90 PI 1 with a higher TMPE-OTA content were selected for the formulation of resins with BAG particles. In addition, though the dye content showed no direct effect on photopolymerization of the monomers, it clearly reduced the overcuring. Therefore, the quantity of dye in the following experiments was related to the mass fraction of BAG, i.e., the higher the percentage of BAG, the higher the amount of dye required to control the light scattering.

Optimization of Ceramic Resin Compositions
To investigate the influence of the BAG content on the viscosity of the ceramic resins, xBM 90 PI 1 D 0.005 resins with 1% PI, 0.005% dye and different BAG contents, ranging from 20 to 60% (m/m resin ), were compared (Table 2). PEG-200 was not incorporated in any of these resins. The viscosity of resin compositions 20BM 90 PI 1 D 0.005 , 30BM 90 PI 1 D 0.005 and 40BM 90 PI 1 D 0.005 was very low ( Figure 5A).The resin 20BM 90 PI 1 D 0.005 showed non-Newtonian viscosity ( Figure 5B) at low shear rates due to inhomogeneous dispersion at lower BAG contents, while both 30BM 90 PI 1 D 0.005 and 40BM 90 PI 1 D 0.005 resins showed shear thinning behavior at shear rates below 30 s −1 . At higher BAG contents, the viscosity of 50BM 90 PI 1 D 0.005 and 60BM 90 PI 1 D 0.005 was very high (up to 50 Pa.s) at low shear rates but decreased with an increase in shear rate. It is a typical shear thinning behavior (pseudoplastic) of glass/ceramic inks. In the next step, compositions with a higher BAG content, i.e., 45-60% were selected for further optimization (i.e., stability, hardness, and rate of photopolymerization) of the photocurable resin.
To improve the rheological behavior of BAG containing resins, PEG-200 was employed. PEG has been reported to improve resin flowability while being non-reactive [24]. Figure 6A shows the rheology of resins 60BM90PI1D0.005Ee with 60% m/mresin BAG and varying amounts (i.e., 0, 10 and 15%) of PEG-200. The viscosity of the resin 60BM90PI1D0.005E0 was 14.9 Pa.s at a shear rate of 20 s −1 , while on addition of 10% PEG-200, the viscosity value decreased to 6.6 Pa.s. On further increasing the PEG-200 contents to 15%, the viscosity value decreased to 4.2 Pa.s. However, the difference in viscosity for 10 and 15% PEG-200 was small. Therefore, 10% mass fraction of PEG-200 was selected for further experiments. Figure 6B shows the influence of BAG contents on the rheology of xBM90PI1D0.005E10 with 10% PEG-200. In presence of PEG-200, the rheology was almost independent of the amount of BAG (55% to 65% mass fraction) [23], which allows the application of these bioactive glass resins with high filler content in stereolithography. In the next step, compositions with a higher BAG content, i.e., 45-60% were selected for further optimization (i.e., stability, hardness, and rate of photopolymerization) of the photocurable resin.
To improve the rheological behavior of BAG containing resins, PEG-200 was employed. PEG has been reported to improve resin flowability while being non-reactive [24]. Figure 6A shows the rheology of resins 60BM 90 PI 1 D 0.005 E e with 60% m/m resin BAG and varying amounts (i.e., 0, 10 and 15%) of PEG-200. The viscosity of the resin 60BM 90 PI 1 D 0.005 E 0 was 14.9 Pa.s at a shear rate of 20 s −1 , while on addition of 10% PEG-200, the viscosity value decreased to 6.6 Pa.s. On further increasing the PEG-200 contents to 15%, the viscosity value decreased to 4.2 Pa.s. However, the difference in viscosity for 10 and 15% PEG-200 was small. Therefore, 10% mass fraction of PEG-200 was selected for further experiments. Figure 6B shows the influence of BAG contents on the rheology of xBM 90 PI 1 D 0.005 E 10 with 10% PEG-200. In presence of PEG-200, the rheology was almost independent of the amount of BAG (55% to 65% mass fraction) [23], which allows the application of these bioactive glass resins with high filler content in stereolithography.  Figure 7. Resin compositions 55BM80PI1D0.005E0 started to settle on day 2, as can be seen in Figure 7B, while resin composition 55BM80PI1D0.005E10 was stable even after day 43. Hence, PEG-200 improved the stability and shelf life of the resin. A stable and homogeneous suspension leads to the homogeneity of the microstructure of the sintered parts and improved mechanical strength.  Figure 7. Resin compositions 55BM 80 PI 1 D 0.005 E 0 started to settle on day 2, as can be seen in Figure 7B, while resin composition 55BM 80 PI 1 D 0.005 E 10 was stable even after day 43. Hence, PEG-200 improved the stability and shelf life of the resin. A stable and homogeneous suspension leads to the homogeneity of the microstructure of the sintered parts and improved mechanical strength.  The effect of the amount of PEG-200 on the shore-D hardness of samples (curing time = 60 s) was also investigated ( Table 3). The shore-D measurements showed that the hardness value decreases with increasing amounts of PEG-200. This is because PEG-200 acts as a plasticizer and is also known to lowers the Young's modulus [25]. However, when increasing the weight fraction of PEG-200 from 10 to 15%, a very small change in rheology was observed, as discussed above. Henceforth, a 10% weight fraction of PEG-200 was chosen for the additive manufacturing of BAG resin compositions in order to avoid any change in mechanical properties. Table 3. Shore-D data of photocured samples (n = 5) in relation to PEG-200 content for compositions The effect of the amount of PEG-200 on the shore-D hardness of samples (curing time = 60 s) was also investigated ( Table 3). The shore-D measurements showed that the hardness value decreases with increasing amounts of PEG-200. This is because PEG-200 acts as a plasticizer and is also known to lowers the Young's modulus [25]. However, when increasing the weight fraction of PEG-200 from 10 to 15%, a very small change in rheology was observed, as discussed above. Henceforth, a 10% weight fraction of PEG-200 was chosen for the additive manufacturing of BAG resin compositions in order to avoid any change in mechanical properties. Furthermore, the influence of the BAG content on the photopolymerization reaction was investigated by in situ UV-DSC. Figure 8 shows the effect of different amounts of BAG on the photopolymerization reaction of photocurable resins with 1% PI and 0.01% dye (both m/m acr ) and 10% PEG-200 (m/m resin ). The thermogram indicates that the photopolymerization reaction for resin compositions with BAG was unfinished even after 3 min of measurement. These BAG resin compositions xBM 80 PI 1 D 0.01 E 10 contain a high mass fraction of triacrylate (80TMPE-OTA:20HEA), and therefore, these mixtures would tend to trap more radicals and more unreacted monomer. Another factor might be the scattering effect and absorption of light by BAG particles due to the difference in refractive index value of monomers and fillers. The refractive index of BAG (bioglass 45S5) is 1.54-1.56 [21], while the refractive index of acrylates is 1.44-1.47. Figure 8 also shows that there was a difference in heat profile in relation to the BAG content in the resins. At high BAG loading, less heat was released per unit mass. An increase in weight fraction of BAG per unit mass in the composition leads to a relative decrease in the number of monomers and, consequently, less heat is released per gram. In the case of 40BM 80 PI 1 D 0.01 E 10 , two endothermic dips (fluctuations) were observed, one after 0.8 min and the other at 1.5 min, which could be related to inhomogeneity of the resin with lower BAG contents. Hence, BAG contents not only affect the rheology of the resin but also strongly influence the rate of the polymerization reaction.
Biomedicines 2022, 10, x FOR PEER REVIEW 12 of 18 mass in the composition leads to a relative decrease in the number of monomers and, consequently, less heat is released per gram. In the case of 40BM80PI1D0.01E10, two endothermic dips (fluctuations) were observed, one after 0.8 min and the other at 1.5 min, which could be related to inhomogeneity of the resin with lower BAG contents. Hence, BAG contents not only affect the rheology of the resin but also strongly influence the rate of the polymerization reaction. As discussed above, BAG particles scatter light due to a refractive index contrast with the monomer and, consequently, the cure depth is affected, which can cause overgrowth or uncontrolled curing [26]. The dye has been reported to suppress light scattering (due to deflection by filler particles) and control the overgrowth as mentioned earlier [27]. In situ UV-DSC measurements of BAG resins showed difference in heat released per unit mass for different amounts of the dye in the resin ( Figure 9A). For the resin 50BM90PI1D0 without dye (a), the heat released was about two times greater than for resin 50BM90PI1D0.01 (b). Hence, the dye has a direct correlation with the presence of BAG. It absorbs the free radicals and prevents the protuberant scattering effect of BAG particles. This is supported by Figure 9B, where the cure depth of two different resin compositions As discussed above, BAG particles scatter light due to a refractive index contrast with the monomer and, consequently, the cure depth is affected, which can cause overgrowth or uncontrolled curing [26]. The dye has been reported to suppress light scattering (due to deflection by filler particles) and control the overgrowth as mentioned earlier [27]. In situ UV-DSC measurements of BAG resins showed difference in heat released per unit mass for different amounts of the dye in the resin ( Figure 9A). For the resin 50BM 90 PI 1 D 0 without dye (a), the heat released was about two times greater than for resin 50BM 90 PI 1 D 0.01 (b). Hence, the dye has a direct correlation with the presence of BAG. It absorbs the free radicals and prevents the protuberant scattering effect of BAG particles. This is supported by Figure 9B, where the cure depth of two different resin compositions 50BM 90 PI 1 D 0.005 and 50BM 90 PI 1 D 0.01 with various amounts of dye were investigated. The cure depth of sample 50BM 90 PI 1 D 0.005 was 0.4 mm after 30 s light exposure time, while no curing was observed for 50BM 90 PI 1 D 0.01 . Hence, an increase in amount of dye from 0.005 to 0.01% (m/m acr ) delayed the curing time. In order to investigate the effect of PI on the photopolymerization, two different resin compositions, 55BM80PI1D0.01E10 and 55BM80PI2D0.01E10, with 1 and 2% (m/macr) PI, respectively, were selected and their effect on photo-curing was investigated by UV-DSC. Figure  10 shows expeditious curing of the resin with higher percentage of PI due to the availability of more free radicals. Nevertheless, the resin composition with 1% m/macr PI was selected for all photo-curing reactions to avoid rapid conversion, which can cause crack formation. In order to investigate the effect of PI on the photopolymerization, two different resin compositions, 55BM 80 PI 1 D 0.01 E 10 and 55BM 80 PI 2 D 0.01 E 10 , with 1 and 2% (m/m acr ) PI, respectively, were selected and their effect on photo-curing was investigated by UV-DSC. Figure 10 shows expeditious curing of the resin with higher percentage of PI due to the availability of more free radicals. Nevertheless, the resin composition with 1% m/m acr PI was selected for all photo-curing reactions to avoid rapid conversion, which can cause crack formation.

Fabrication of 3D Scaffolds
Two different ceramic resin compositions, 55BM80PI1D0.015E10 and 60BM80PI1D0.015E10, were chosen for 3D printing in the Lithoz Ceracet. The layer thickness was set at 25 µ m

Fabrication of 3D Scaffolds
Two different ceramic resin compositions, 55BM 80 PI 1 D 0.015 E 10 and 60BM 80 PI 1 D 0.015 E 10 , were chosen for 3D printing in the Lithoz Ceracet. The layer thickness was set at 25 µm with backlight exposure of 4 s and main light exposure of 4 s for each layer. In order to investigate the effect of BAG loading, first a gear-shaped sample was prepared with two different dimensions. The printing test was completed successfully with an accurate desired structure. Moreover, no scattering effect was observed with these two compositions. On the macro-scale, the parts appeared to be defect-free and the surface was flat and smooth.
The µ-CT of the printed samples was performed to investigate the detailed structure and homogeneity. As shown in the top view in Figure 11, the printing direction is from bottom to top, which is also expressed by the sequence of µ-CT pictures. The bright area represents the materials and the dark area represents air; the brighter the area, the higher the density of the material. The side view µ-CT images show the layer connections in the printed samples. Comparing the side view images, it appears that the sample 55BM 80 PI 1 D 0.015 E 10 (small gear-shaped structure) is more uniform and homogeneous as compared to the sample 60BM 80 PI 1 D 0.015 E 10 (large gear-shaped structure) with higher bioactive glass loading (60% weight fraction BAG). The layers were interdiffused in 55BM 80 PI 1 D 0.015 E 10 ; however, some grains appeared during µ-CT, which might be due to aggregates. In contrast in 60BM 80 PI 1 D 0.015 E 10 , the delamination of layers was observed at a few places without the appearance of any aggregation.  Figure 12 shows the 3D printed gear-shaped and honey-comb structure with composition 55BM80PI1D0.015E10. No visible cracks or layer separation on the surface of the printed structure was seen in SEM images ( Figure 12B,E). However, a few lines parallel distributed on the surface of the cross-section were observed ( Figure 12C), which can be attributed to the layers' connection in the printing process. The images at various magnifications showed that the printed samples were crack-free and dense (Supplementary information Figure S1).  Figure 12 shows the 3D printed gear-shaped and honey-comb structure with composition 55BM 80 PI 1 D 0.015 E 10 . No visible cracks or layer separation on the surface of the printed structure was seen in SEM images ( Figure 12B,E). However, a few lines parallel distributed on the surface of the cross-section were observed ( Figure 12C), which can be attributed to the layers' connection in the printing process. The images at various magnifications showed that the printed samples were crack-free and dense (Supplementary information Figure S1). SEM and µ -CT results showed that the sample 55BM80PI1D0.015E10 was the most optimized BAG resin composition for ceramic stereolithography of bioactive glass scaffolds. No cracks were observed in printed samples and SEM images at high magnification showed that the printed sample has a dense structure.
Henceforth, a suitable resin composition for 3D printing of bioactive glass was achieved by investigating the physical properties, including rheology, heat flow and morphology. These investigations will be helpful to formulate the resin for 3D printing of ceramics/glasses for various applications.

Conclusions
Ceramic stereolithography is a very suitable method for the fabrication of implants and scaffolds for tissue engineering with controlled porosity as they can be fabricated according to patient-specific requirements. The results showed that the rate of photopolymerization increases with an increase in the concentration of acrylate monomers due to an increase in the number of unsaturated C=C groups per unit gram of monomer units. Resin compositions with a TMPE-OTA:HEA weight ratio of 90:10 and 80:20 were suitable for BAG printing due to optimal resin viscosity. To increase the flowability of BAG resin, PEG-200 was used as diluent and plasticizer, 10% of PEG-200 was found to be the appropriate amount for the fabrication. At this PEG-200 content, the viscosity of bioactive glass resin remains almost constant irrespective of glass loading in the range from 55 to 65 wt.%. The light scattering effect (due to deflection by filler particles) over the slurry layers can be controlled by using appropriate amounts of light absorbent (dye) which absorb photons (without creating free radicals) and help to control the overgrowth. Two compositions, 55BM80PI1D0.015E10 and 60BM80PI1D0.015E10, with 55 and 60% BAG, 10% PEG-200, 1% of PI and 0.015% of Sudan orange G dye were found to be the suitable for the stereolithographic fabrication and were successfully used to fabricate uniform scaffolds with different shapes without the appearance of macro-cracks.
Supplementary Materials: The following supporting information can be downloaded at: www.mdpi.com/xxx/s1, Tables S1 (Properties of the applied acrylate monomers and polymers as given by suppliers), Tables S2 (The compositions of the resins without BAG investigated in this study), Tables S3 (The compositions of the bioactive glass containing resins investigated in this study) and Figure S1 (The SEM images of the 3D printed samples showing the absence of cracks). SEM and µ-CT results showed that the sample 55BM 80 PI 1 D 0.015 E 10 was the most optimized BAG resin composition for ceramic stereolithography of bioactive glass scaffolds. No cracks were observed in printed samples and SEM images at high magnification showed that the printed sample has a dense structure.
Henceforth, a suitable resin composition for 3D printing of bioactive glass was achieved by investigating the physical properties, including rheology, heat flow and morphology. These investigations will be helpful to formulate the resin for 3D printing of ceramics/glasses for various applications.

Conclusions
Ceramic stereolithography is a very suitable method for the fabrication of implants and scaffolds for tissue engineering with controlled porosity as they can be fabricated according to patient-specific requirements. The results showed that the rate of photopolymerization increases with an increase in the concentration of acrylate monomers due to an increase in the number of unsaturated C=C groups per unit gram of monomer units. Resin compositions with a TMPE-OTA:HEA weight ratio of 90:10 and 80:20 were suitable for BAG printing due to optimal resin viscosity. To increase the flowability of BAG resin, PEG-200 was used as diluent and plasticizer, 10% of PEG-200 was found to be the appropriate amount for the fabrication. At this PEG-200 content, the viscosity of bioactive glass resin remains almost constant irrespective of glass loading in the range from 55 to 65 wt.%. The light scattering effect (due to deflection by filler particles) over the slurry layers can be controlled by using appropriate amounts of light absorbent (dye) which absorb photons (without creating free radicals) and help to control the overgrowth. Two compositions, 55BM 80 PI 1 D 0.015 E 10 and 60BM 80 PI 1 D 0.015 E 10 , with 55 and 60% BAG, 10% PEG-200, 1% of PI and 0.015% of Sudan orange G dye were found to be the suitable for the stereolithographic fabrication and were successfully used to fabricate uniform scaffolds with different shapes without the appearance of macro-cracks.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/biomedicines10020395/s1, Table S1 (Properties of the applied acrylate monomers and polymers as given by suppliers), Table S2 (The compositions of the resins without BAG investigated in this study), Table S3 (The compositions of the bioactive glass containing resins investigated in this study) and Figure S1