- Email: [email protected]
Digital light processing of wollastonite-diopside glass-ceramic complex structures
Accepted Manuscript Title: Digital Light Processing of Wollastonite-Diopside Glass-ceramic Complex Structures Authors: Johanna Schmidt, Hamada Elsayed, Enrico Bernardo, Paolo Colombo PII: DOI: Reference:
S0955-2219(18)30366-2 https://doi.org/10.1016/j.jeurceramsoc.2018.06.004 JECS 11925
To appear in:
Journal of the European Ceramic Society
Received date: Revised date: Accepted date:
8-2-2018 30-5-2018 2-6-2018
Please cite this article as: Schmidt J, Elsayed H, Bernardo E, Colombo P, Digital Light Processing of Wollastonite-Diopside Glass-ceramic Complex Structures, Journal of the European Ceramic Society (2018), https://doi.org/10.1016/j.jeurceramsoc.2018.06.004 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Digital Light Processing of Wollastonite-Diopside Glass-ceramic Complex Structures
Johanna Schmidt1, Hamada Elsayed1,2, Enrico Bernardo1, Paolo Colombo1,3,* 1
Dipartimento di Ingegneria Industriale, University of Padova, Via Marzolo 9, 35131 Padova, Italy
2
Ceramics Department, National Research Centre, El-Bohous Street, 12622 Cairo, Egypt
3
IP T
Department of Materials Science and Engineering, the Pennsylvania State University, University Park, PA 16801, US
Abstract
A
N
U
SC R
In this work, the possibility of shaping a glass-filled photosensitive polymer resin with Digital Light Processing (DLP) into a complex 3D structure and transforming it subsequently into a bioactive glassceramic scaffold was investigated. The influence of the printing conditions and the heat-treatment was studied using a 41 vol% glass-filled acrylated polymer resin. Scaffolds with designed architecture were turned into a wollastonite-diopside glass-ceramic at 1100 °C. They completely maintained their shape, exhibited no viscous flow and showed a homogenous linear shrinkage of around 25 %. At 83 vol% porosity structures with Kelvin cell design exhibited a compressive strength exceeding 3 MPa, demonstrating that the material is suitable for the fabrication of bioceramic scaffolds for bone tissue engineering applications.
ED
M
Key words: Additive Manufacturing, Digital Light Processing, Bioactive glass-ceramic, Scaffolds, Stereolithography
1. Introduction
A
CC E
PT
Extensive research has been conducted in the past few years in the field of tissue engineering for repairing diseased or damaged tissues. The regeneration of biological tissue can be aided by synthetic materials, so called “scaffolds”, if they are biocompatible, osteoconductive and bioactive [1]. Ideally, they are biodegradable and will be replaced at the same rate at which the extracellular matrix is formed by the cells [2]. Open and interconnected porosity in scaffolds is necessary for the vascularisation and cell ingrowth [3, 4]. Therefore, it is not only necessary to have scaffolds with a high degree of porosity but also with a precisely controlled internal architecture (e.g. geometry, density, pore size and distribution), which should be adjustable to facilitate and optimize the biological properties of the component. Additive manufacturing techniques, especially stereolithography, allow for a precise control over the architecture of produced parts. A simple alteration of a computer model (CAD-file) can be performed to obtain the desired geometric arrangement of struts and pores as well as control their interconnectivity. In stereolithography, a photosensitive resin is polymerized through UV exposure in each printing step and builds up the structure in the z-direction layer-by-layer. In the laser configuration, stereolithography (SLA), a UV laser scans the printing area and polymerizes the liquid oligomers while in digital light processing (DLP) a projector illuminates the complete area at once [5]. Stereolithography of ceramics and glasses is not trivial, due to the fact that, in contrast with pure polymer resins, the light is scattered by the presence of glass and ceramic particles, which can lead to a decrease in resolution
Corresponding author. Email: [email protected]; Phone: +39 049 8275825
[6-8]. Furthermore, the particles have to be stabilized against sedimentation during the printing process to avoid an inhomogeneous distribution of the glass particles inside the printed structure. To achieve a dense glass or ceramic after sintering, a highly filled (40 – 60 vol%) particle suspension has to be employed [6], which usually is achieved by the addition of dispersants to maximize the ceramic loading while ensuring that the viscosity stays low enough for printing [8, 9].
2. Materials and Methods
ED
M
A
N
U
SC R
IP T
In order to avoid these issues, in some researches additive manufacturing was used to fabricate polymer moulds from a photosensitive polymer resin and used as negative moulds, i.e. employing the negative replica technique. They were then filled with the glass or ceramic suspensions via gel-casting to produce porous HA implants [10], polymer/HA composite scaffolds [11], bioactive glass [1] or apatite–mullite glass-ceramic [12] after the removal of the negative moulds during heat-treatment. As the scaffolds were not directly produced by stereolithography, the replica method has limitations regarding the geometries and overall dimensions that can be fabricated [12], as the gels need to be able to penetrate the negatives and the mould has to be completely eliminated after the setting of the positive. Putlyaev et al. showed the direct fabrication of bioactive ceramics via stereolithography of calcium alkali metal double phosphates [13]. However, as calcium phosphate ceramic structures do not usually show the necessary mechanical strength to withstand severe mechanical stress, their use is usually limited to areas where mechanical stability is not crucial [3]. Also, the fabrication of 45S5 Bioglass-structures via stereolithography was shown by Tesavibul et al., but structures exhibited rather low mechanical properties [14]. A new biphasic biomaterial in the CaSiO3-CaMg(SiO3)2 (wollastonitediopside) system was first proposed by Sainz et al. to take advantage of the biological activity of the biomaterial while ensuring enough mechanical stability [3]. We report here for the first time Digital Light Processing of a bioactive wollastonite-diopside glass-ceramic, which shows generally better mechanical properties than bioactive glass [15, 16]. Highly complex structures, from a material which has been proven to be biocompatible [3, 15] can be realized in a time frame appropriate for further scaling up.
CC E
PT
The glass used in the experiments had the composition corresponding to the eutectic in the CaO-MgOSiO2 phase diagram [3]. The glass, produced from pure minerals and chemicals, had the overall composition of 53.98 wt% SiO2, 34.58 wt% CaO and 11.7 wt% MgO, leading to 52 mol% wollastonite (CaO·SiO2) and 48 mol% diopside (CaO·MgO·2SiO2). W-D glass was produced from pure minerals and chemicals (silica, dolomite, calcium carbonate - all in powders <10 µm, Industrie Bitossi, Vinci, Italy), by melting in a platinum crucible at a temperature of 1400°C (heating rate of 10°C/min). After ball milling, grinding and sieving, glass particles with a diameter < 45 µm were kept and used in the printing process.
A
The glass particles were added to an acrylated polymer resin (tripropylene glycol diacrylate, Robofactory, Italy), and homogenously distributed via magnetic stirring overnight. A particle loading of 63 wt% was used, which ensured that the viscosity was low enough for printing without any additional solvent in the printing system. As the resin was developed to be used with the DLP printer we employed in this work (3DLPrinter-HD 2.0, Robofactory, Italy), no additional photoinitiator or photoabsorber had to be added to the system. The bottles were wrapped in aluminium foil throughout the mixing process to eliminate light illumination from the surroundings. The glass filled-polymer resin was poured in a self-made PDMS printing chamber and printed using the visible light radiation (400 – 500 nm) of the DLP printer. A 80 μm layer height was set during printing. After the printing process, the 3D printed structures were cleaned in isopropanol for 3 min in an ultrasonic bath, to remove the uncured resin, and then dried with compressed air. To ensure a full cure of the acrylic network, the
structures were then placed in an UV furnace (365 nm, Robofactory, Italy) and irradiated for an additional 15 min. Finally, the structures were heated with 1°C/min to 500°C with a holding time of 3h to decompose the polymer network and sintered with 5°C/min at 1100°C for 1 h to develop wollastonite-diopside glass-ceramic structures. The heating schedule was selected on the basis of previous experiments and DTA results [15].
U
SC R
IP T
Morphological characterization of the 3D structures was performed by stereo-microscopy (STEMI 2000-C, Zeiss, USA) and by scanning electron microscopy (SEM, Quanta 450, FEI, USA) on metal-coated samples. Thermo-gravimetric analysis (TGA, STA 409/429 Netzsch Gerätebau GmbH, Selb, Germany) was carried out with printed samples at a heating rate of 5°C/min in air. The crystalline phases were identified by means of X-ray diffraction on powdered samples (XRD; Bruker D8 Advance, Bruker AXS GmbH, Karlsruhe, Germany), supported by data from PDF-2 database (ICDD-International Centre for Diffraction Data, Newtown Square, PA, USA) and Match! program package (Crystal Impact GbR, Bonn, Germany). The true density of the initial glass particles and the finely milled, heat treated printed samples were measured by means of a helium pycnometer (Micromeritics AccuPyc 1330, Norcross, GA, USA). The compressive strength of cubic Kelvin cell structures (7.5 x 7.5 x 7.5 mm3) was evaluated at room temperature with an Instron 1121 UTM (Instron, Danvers, MA, USA), operating with a crosshead speed of 1 mm/min.
N
3. Results and Discussion
M
A
The printing resin was based on the dispersion of 41 vol% of glass particles inside a photocurable acrylated polymer resin. As the glass particles proved to be stable in the liquid resin during the entire time frame of the printing process (up to 3-4 hours) without noticeable precipitation it was not necessary to employ additional stabilization additives [8]. Also, no additional dispersant was required, as the direct dispersion of the glass particles in the polymer resin enabled to achieve a suitably high particle concentration, comparable to what was reported by other authors [1, 6, 13, 17].
CC E
PT
ED
TGA measurement of printed samples confirmed a ceramic yield of 62.8 wt.% (see Fig.1) which is consistent with the 63 wt.% of glass powder (2.76 g/cm3) added to the photosensitive polymeric resin (1.13 g/cm3). This proved that the powder was homogenously distributed not only inside the printing liquid but also in the printed structures and that no expulsion or precipitation of the glass powder during the printing process took place. The acrylate monomers formed during polymerization a network around the glass particles, entrapping them inside the printed structure in the same quantity as in the liquid.
A
A model cube structure (Fig. 2) served as a template to optimize the exposure time of the glass particle filled resin. Stereomicroscope images of the printed model were taken perpendicularly to the top of the printed samples and showed that a minimum exposure time of 1.5 s was necessary to fully fabricate the horizontal struts, which were as small as 100 µm. According to the different illumination tests, lower exposure times did not provide the necessary illumination to cure the smallest cube features, while higher printing times lead to an overexposure of the structure. Microscope images from the side of the model cube structure (Fig. 2) also showed that, at the selected optimal exposure time, the amount of photoabsorber present in the printing resin proved sufficient to limit the penetration depth of the printing light. No addition of further photoinitiator was necessary, as the exposure time selected was already suitable for the efficient printing of large structures.
IP T SC R U N A M ED PT CC E A Figure 1 Thermal analysis and XRD pattern of a printed and heat-treated sample.
Top view
Side view 1
Side view 2
2.0 s
SC R
Top view of cube structure at different exposure times (scale bars = 0.5 mm) 0.5 s 1.0 s 1.5 s
IP T
CAD model scaffold
U
Figure 2 Display of model cube structure (scale bars = 2 mm).
M
A
N
To ensure that the selected heating schedule would yield the desired phase assemblage in the bioceramic component, printed samples were crushed after sintering and analysed by XRD (Fig. 1). The samples yielded pure diopside (CaMgSi2O6) and wollastonite (CaSiO3) crystalline phases after sintering at 1100°C. No additional phase was detected in the heat-treated samples, showing a complete conversion into wollastonite-diopside biocompatible glass-ceramic.
PT
ED
The model cube structures, shown in Fig. 2, were also used to investigate the shrinkage of the wollastonite-diopside glass-ceramic printed samples during heating. A completely homogenous shrinkage in all directions was measured (linear shrinkage of 25.4 ± 1.6 %). TGA analysis (see Fig. 1) indicated that complete burn out of the photocured polymer occurred at 500°C. This means that during the polymer decomposition the glass particles kept the original shape and the green-network was strong enough to support the printed structure without collapsing. Furthermore, the efficient crystallization of the selected glass composition hindered any viscous collapse during the formation of the glass-ceramic, so that a complete shape retention was observed after firing.
A
CC E
After the optimization of printing parameters and heat treatment, unique complex structures, that could be manufactured solely by stereolithography, were printed. We show here samples possessing a Kelvin cell (a tetrakaidecahedron) structure, as an example of the architectural complexity that can be realized using stereolithography. Fig. 3 shows SEM images of the as printed and sintered samples. We can observe that all the features of the as printed sample, including the presence of discrete layers in z-direction, were completely transferred in the sintered sample. The images confirm that the uniform and isotropic shrinkage led to no deformation of the structures after the heat treatment.
IP T SC R U N A M
ED
Figure 3: SEM images of samples with a Kelvin cell architecture: as printed (A, B); sintered at 1100°C (C, D).
A
CC E
PT
The investigation of the fracture surface of a sintered glass-ceramic 3D printed sample (Fig. 4) showed that no internal nor surface microcracks could be found in the structure, indicating that the elimination of the photocurable polymer did not induce any macroscopic defects. Furthermore, it is worth observing that no delamination between the printed layers occurred, and that the fracture did not follow the path of the printed layers. This demonstrates that the interface between two layers was not the week point in the structure and that the individual layers were strongly connected. Finally, while individual layers could be observed on the surface of the printed and of the fired samples, no interfaces inside the structures were visible, suggesting that the viscous flow sintering and crystallization led to a sample possessing a homogeneous microstructure throughout its volume.
Figure 4. SEM image of the fracture surface of a sintered glass-ceramic printed sample.
IP T
The printed wollastonite-diopside glass-ceramic samples, after the heat treatment at 1100°C, possessed a bulk density of 0.47 g/cm3, true density of 3.12 g/cm3, and total porosity of 83 vol%. The apparent and true density were measured for printed glass-ceramic sample without base after sintering (Fig. 3C) and showed approximately the same value. Therefore, the struts have not closed porosity. Some amount of residual porosity could be observed on the surface of the sample as well as on the fracture surface. This could be attributed to the crystallization and transition into a glassceramic material, although the contribution of incomplete viscous flow and sintering to its presence should also be considered. However, it should be noted that this porosity in the struts, which is typical for a diopside and wollastonite glass-ceramic [18], can be useful for biological applications as a large surface area is beneficial for cell attachment and spreading [19, 20].
N
U
SC R
The cubic Kelvin cell structures (represented in Fig. 3) without base show a mechanical strength of 3.2 ± 0.5 MPa at a total porosity of 83 vol%. This demonstrates that the properties of the developed wollastonite-diopside glass-ceramic structures are within the standard reference range for the mechanical strength (2–12 MPa) [21] as well as for the porosity (75-95 vol%) [22] typical of human trabecular bone. The crushing strength is excellent, if we consider the classical model proposed by Gibson and Ashby [23]. For a bending-dominated structure, like the present one, the crushing strength (σc) depends on the bending strength (σbend) of the solid phase and on the relative density (ρrel) according to an exponential correlation: σc= σbend C ρrel3/2, where C is a dimensionless constant (being ≈0.2). From experimental value of σc and the low relative density (ρrel=0.17, for a total porosity of 83%) we can infer that the strength of the solid phase well exceeds 200 MPa.
A
4. Conclusions
CC E
PT
ED
M
Bioactive glass-ceramic samples possessing a highly complex architecture were fabricated using a photocurable polymer containing a high volume of glass particles, which crystallized into wollastonite and diopside upon heat-treatment. The glass particles were trapped in the cured polymer and highly porous structures were built up layer-by-layer via DLP, after optimization of the printing parameters. The heat-treatment in air, at 1100°C, caused the burn-out of the polymer and the sinter-crystallization of glass particles. While the material exhibited a residual porosity due to the formation of crystals, no delamination and microcrack formation could be observed in the structure. The samples showed a homogeneous shrinkage with complete shape retention. The bioactivity of wollastonite-diopside glassceramics, combined with the design freedom and high resolution of 3D stereolithography, as well as the good mechanical strength make these structures suitable candidates for bone tissue engineering applications.
Acknowledgments
A
J.S. gratefully acknowledges the CARIPARO Foundation, Padova, Italy, for her PhD scholarship. References
[1] S. Padilla, S. Sánchez-Salcedo, M. Vallet-Regí, Bioactive glass as precursor of designedarchitecture scaffolds for tissue engineering, Journal of Biomedical Materials Research Part A 81A(1) (2007) 224-232. [2] T.M. Freyman, I.V. Yannas, L.J. Gibson, Cellular materials as porous scaffolds for tissue engineering, Progress in Materials Science 46(3) (2001) 273-282.
A
CC E
PT
ED
M
A
N
U
SC R
IP T
[3] M.A. Sainz, P. Pena, S. Serena, A. Caballero, Influence of design on bioactivity of novel CaSiO3–CaMg(SiO3)2 bioceramics: In vitro simulated body fluid test and thermodynamic simulation, Acta Biomaterialia 6(7) (2010) 2797-2807. [4] B. Flautre, M. Descamps, C. Delecourt, M.C. Blary, P. Hardouin, Porous HA ceramic for bone replacement: Role of the pores and interconnections – experimental study in the rabbit, Journal of Materials Science: Materials in Medicine 12(8) (2001) 679-682. [5] A. Zocca, P. Colombo, C.M. Gomes, J. Günster, Additive Manufacturing of Ceramics: Issues, Potentialities, and Opportunities, Journal of the American Ceramic Society 98(7) (2015) 1983-2001. [6] N. Travitzky, A. Bonet, B. Dermeik, T. Fey, I. Filbert-Demut, L. Schlier, T. Schlordt, P. Greil, Additive Manufacturing of Ceramic-Based Materials, Advanced Engineering Materials 16(6) (2014) 729-754. [7] J.H. Jang, S. Wang, S.M. Pilgrim, W.A. Schulze, Preparation and characterization of barium titanate suspensions for stereolithography, Journal of the American Ceramic Society 83(7) (2000) 1804-1806. [8] J.W. Halloran, V. Tomeckova, S. Gentry, S. Das, P. Cilino, D. Yuan, R. Guo, A. Rudraraju, P. Shao, T. Wu, T.R. Alabi, W. Baker, D. Legdzina, D. Wolski, W.R. Zimbeck, D. Long, Photopolymerization of powder suspensions for shaping ceramics, Journal of the European Ceramic Society 31(14) (2011) 2613-2619. [9] Y. De Hazan, J. Heinecke, A. Weber, T. Graule, High solids loading ceramic colloidal dispersions in UV curable media via comb-polyelectrolyte surfactants, Journal of Colloid and Interface Science 337(1) (2009) 66-74. [10] T.-M. Chu, J. Halloran, S. Hollister, S. E. Feinberg, Hydroxyapatite Implants with Designed Internal Architecture, 2001. [11] S. Limpanuphap, B. Derby, Manufacture of biomaterials by a novel printing process, Journal of Materials Science: Materials in Medicine 13(12) (2002) 1163-1166. [12] K. Chopra, P. Mummery, B. Derby, J. Gough, Gel-cast glass-ceramic tissue scaffolds of controlled architecture produced via stereolithography of moulds, 2012. [13] V.I. Putlyaev, P.V. Evdokimov, T.V. Safronova, E.S. Klimashina, N.K. Orlov, Fabrication of osteoconductive Ca3-x M-2x (PO4)(2) (M = Na, K) calcium phosphate bioceramics by stereolithographic 3D printing, Inorg. Mater. 53(5) (2017) 529-535. [14] P. Tesavibul, R. Felzmann, S. Gruber, R. Liska, I. Thompson, A.R. Boccaccini, J. Stampfl, Processing of 45S5 Bioglass® by lithography-based additive manufacturing, Materials Letters 74 (2012) 81-84. [15] H. Elsayed, A. Rincón Romero, L. Ferroni, C. Gardin, B. Zavan, E. Bernardo, Bioactive Glass-Ceramic Scaffolds from Novel ‘Inorganic Gel Casting’ and Sinter-Crystallization, Materials 10(2) (2017) 171. [16] F. Baino, C. Vitale-Brovarone, Mechanical properties and reliability of glass–ceramic foam scaffolds for bone repair, Materials Letters 118 (2014) 27-30. [17] S. Zhang, N. Sha, Z. Zhao, Surface modification of α-Al2O3 with dicarboxylic acids for the preparation of UV-curable ceramic suspensions, 2016. [18] A. Karamanov, M. Pelino, Induced Crystallization Porosity and Properties of Sintereds Diopside and Wollastonite glass–ceramics, 2008. [19] O. Chan, M.J. Coathup, A. Nesbitt, C.Y. Ho, K.A. Hing, T. Buckland, C. Campion, G.W. Blunn, The effects of microporosity on osteoinduction of calcium phosphate bone graft substitute biomaterials, Acta Biomaterialia 8(7) (2012) 2788-2794. [20] A. Bignon, J. Chouteau, J. Chevalier, G. Fantozzi, J.P. Carret, P. Chavassieux, G. Boivin, M. Melin, D. Hartmann, Effect of Micro- and Macroporosity of bone substitutes on their mechanical properties and cellular response, 2004. [21] L.L. Hench, J. Wilson, An Introduction to Bioceramics, 2nd ed., Imperial College Press, 1993.
A
CC E
PT
ED
M
A
N
U
SC R
IP T
[22] M.A. Meyers, P.-Y. Chen, Biological Materials Science: Biological Materials, Bioinspired Materials, and Biomaterials, Cambridge University Press, England, 2014. [23] L.J. Gibson, M.F. Ashby, Cellular Solids: Structure and Properties, 2nd ed., Cambridge University Press, Cambridge, 1997.
S0955-2219(18)30366-2 https://doi.org/10.1016/j.jeurceramsoc.2018.06.004 JECS 11925
To appear in:
Journal of the European Ceramic Society
Received date: Revised date: Accepted date:
8-2-2018 30-5-2018 2-6-2018
Please cite this article as: Schmidt J, Elsayed H, Bernardo E, Colombo P, Digital Light Processing of Wollastonite-Diopside Glass-ceramic Complex Structures, Journal of the European Ceramic Society (2018), https://doi.org/10.1016/j.jeurceramsoc.2018.06.004 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Digital Light Processing of Wollastonite-Diopside Glass-ceramic Complex Structures
Johanna Schmidt1, Hamada Elsayed1,2, Enrico Bernardo1, Paolo Colombo1,3,* 1
Dipartimento di Ingegneria Industriale, University of Padova, Via Marzolo 9, 35131 Padova, Italy
2
Ceramics Department, National Research Centre, El-Bohous Street, 12622 Cairo, Egypt
3
IP T
Department of Materials Science and Engineering, the Pennsylvania State University, University Park, PA 16801, US
Abstract
A
N
U
SC R
In this work, the possibility of shaping a glass-filled photosensitive polymer resin with Digital Light Processing (DLP) into a complex 3D structure and transforming it subsequently into a bioactive glassceramic scaffold was investigated. The influence of the printing conditions and the heat-treatment was studied using a 41 vol% glass-filled acrylated polymer resin. Scaffolds with designed architecture were turned into a wollastonite-diopside glass-ceramic at 1100 °C. They completely maintained their shape, exhibited no viscous flow and showed a homogenous linear shrinkage of around 25 %. At 83 vol% porosity structures with Kelvin cell design exhibited a compressive strength exceeding 3 MPa, demonstrating that the material is suitable for the fabrication of bioceramic scaffolds for bone tissue engineering applications.
ED
M
Key words: Additive Manufacturing, Digital Light Processing, Bioactive glass-ceramic, Scaffolds, Stereolithography
1. Introduction
A
CC E
PT
Extensive research has been conducted in the past few years in the field of tissue engineering for repairing diseased or damaged tissues. The regeneration of biological tissue can be aided by synthetic materials, so called “scaffolds”, if they are biocompatible, osteoconductive and bioactive [1]. Ideally, they are biodegradable and will be replaced at the same rate at which the extracellular matrix is formed by the cells [2]. Open and interconnected porosity in scaffolds is necessary for the vascularisation and cell ingrowth [3, 4]. Therefore, it is not only necessary to have scaffolds with a high degree of porosity but also with a precisely controlled internal architecture (e.g. geometry, density, pore size and distribution), which should be adjustable to facilitate and optimize the biological properties of the component. Additive manufacturing techniques, especially stereolithography, allow for a precise control over the architecture of produced parts. A simple alteration of a computer model (CAD-file) can be performed to obtain the desired geometric arrangement of struts and pores as well as control their interconnectivity. In stereolithography, a photosensitive resin is polymerized through UV exposure in each printing step and builds up the structure in the z-direction layer-by-layer. In the laser configuration, stereolithography (SLA), a UV laser scans the printing area and polymerizes the liquid oligomers while in digital light processing (DLP) a projector illuminates the complete area at once [5]. Stereolithography of ceramics and glasses is not trivial, due to the fact that, in contrast with pure polymer resins, the light is scattered by the presence of glass and ceramic particles, which can lead to a decrease in resolution
Corresponding author. Email: [email protected]; Phone: +39 049 8275825
[6-8]. Furthermore, the particles have to be stabilized against sedimentation during the printing process to avoid an inhomogeneous distribution of the glass particles inside the printed structure. To achieve a dense glass or ceramic after sintering, a highly filled (40 – 60 vol%) particle suspension has to be employed [6], which usually is achieved by the addition of dispersants to maximize the ceramic loading while ensuring that the viscosity stays low enough for printing [8, 9].
2. Materials and Methods
ED
M
A
N
U
SC R
IP T
In order to avoid these issues, in some researches additive manufacturing was used to fabricate polymer moulds from a photosensitive polymer resin and used as negative moulds, i.e. employing the negative replica technique. They were then filled with the glass or ceramic suspensions via gel-casting to produce porous HA implants [10], polymer/HA composite scaffolds [11], bioactive glass [1] or apatite–mullite glass-ceramic [12] after the removal of the negative moulds during heat-treatment. As the scaffolds were not directly produced by stereolithography, the replica method has limitations regarding the geometries and overall dimensions that can be fabricated [12], as the gels need to be able to penetrate the negatives and the mould has to be completely eliminated after the setting of the positive. Putlyaev et al. showed the direct fabrication of bioactive ceramics via stereolithography of calcium alkali metal double phosphates [13]. However, as calcium phosphate ceramic structures do not usually show the necessary mechanical strength to withstand severe mechanical stress, their use is usually limited to areas where mechanical stability is not crucial [3]. Also, the fabrication of 45S5 Bioglass-structures via stereolithography was shown by Tesavibul et al., but structures exhibited rather low mechanical properties [14]. A new biphasic biomaterial in the CaSiO3-CaMg(SiO3)2 (wollastonitediopside) system was first proposed by Sainz et al. to take advantage of the biological activity of the biomaterial while ensuring enough mechanical stability [3]. We report here for the first time Digital Light Processing of a bioactive wollastonite-diopside glass-ceramic, which shows generally better mechanical properties than bioactive glass [15, 16]. Highly complex structures, from a material which has been proven to be biocompatible [3, 15] can be realized in a time frame appropriate for further scaling up.
CC E
PT
The glass used in the experiments had the composition corresponding to the eutectic in the CaO-MgOSiO2 phase diagram [3]. The glass, produced from pure minerals and chemicals, had the overall composition of 53.98 wt% SiO2, 34.58 wt% CaO and 11.7 wt% MgO, leading to 52 mol% wollastonite (CaO·SiO2) and 48 mol% diopside (CaO·MgO·2SiO2). W-D glass was produced from pure minerals and chemicals (silica, dolomite, calcium carbonate - all in powders <10 µm, Industrie Bitossi, Vinci, Italy), by melting in a platinum crucible at a temperature of 1400°C (heating rate of 10°C/min). After ball milling, grinding and sieving, glass particles with a diameter < 45 µm were kept and used in the printing process.
A
The glass particles were added to an acrylated polymer resin (tripropylene glycol diacrylate, Robofactory, Italy), and homogenously distributed via magnetic stirring overnight. A particle loading of 63 wt% was used, which ensured that the viscosity was low enough for printing without any additional solvent in the printing system. As the resin was developed to be used with the DLP printer we employed in this work (3DLPrinter-HD 2.0, Robofactory, Italy), no additional photoinitiator or photoabsorber had to be added to the system. The bottles were wrapped in aluminium foil throughout the mixing process to eliminate light illumination from the surroundings. The glass filled-polymer resin was poured in a self-made PDMS printing chamber and printed using the visible light radiation (400 – 500 nm) of the DLP printer. A 80 μm layer height was set during printing. After the printing process, the 3D printed structures were cleaned in isopropanol for 3 min in an ultrasonic bath, to remove the uncured resin, and then dried with compressed air. To ensure a full cure of the acrylic network, the
structures were then placed in an UV furnace (365 nm, Robofactory, Italy) and irradiated for an additional 15 min. Finally, the structures were heated with 1°C/min to 500°C with a holding time of 3h to decompose the polymer network and sintered with 5°C/min at 1100°C for 1 h to develop wollastonite-diopside glass-ceramic structures. The heating schedule was selected on the basis of previous experiments and DTA results [15].
U
SC R
IP T
Morphological characterization of the 3D structures was performed by stereo-microscopy (STEMI 2000-C, Zeiss, USA) and by scanning electron microscopy (SEM, Quanta 450, FEI, USA) on metal-coated samples. Thermo-gravimetric analysis (TGA, STA 409/429 Netzsch Gerätebau GmbH, Selb, Germany) was carried out with printed samples at a heating rate of 5°C/min in air. The crystalline phases were identified by means of X-ray diffraction on powdered samples (XRD; Bruker D8 Advance, Bruker AXS GmbH, Karlsruhe, Germany), supported by data from PDF-2 database (ICDD-International Centre for Diffraction Data, Newtown Square, PA, USA) and Match! program package (Crystal Impact GbR, Bonn, Germany). The true density of the initial glass particles and the finely milled, heat treated printed samples were measured by means of a helium pycnometer (Micromeritics AccuPyc 1330, Norcross, GA, USA). The compressive strength of cubic Kelvin cell structures (7.5 x 7.5 x 7.5 mm3) was evaluated at room temperature with an Instron 1121 UTM (Instron, Danvers, MA, USA), operating with a crosshead speed of 1 mm/min.
N
3. Results and Discussion
M
A
The printing resin was based on the dispersion of 41 vol% of glass particles inside a photocurable acrylated polymer resin. As the glass particles proved to be stable in the liquid resin during the entire time frame of the printing process (up to 3-4 hours) without noticeable precipitation it was not necessary to employ additional stabilization additives [8]. Also, no additional dispersant was required, as the direct dispersion of the glass particles in the polymer resin enabled to achieve a suitably high particle concentration, comparable to what was reported by other authors [1, 6, 13, 17].
CC E
PT
ED
TGA measurement of printed samples confirmed a ceramic yield of 62.8 wt.% (see Fig.1) which is consistent with the 63 wt.% of glass powder (2.76 g/cm3) added to the photosensitive polymeric resin (1.13 g/cm3). This proved that the powder was homogenously distributed not only inside the printing liquid but also in the printed structures and that no expulsion or precipitation of the glass powder during the printing process took place. The acrylate monomers formed during polymerization a network around the glass particles, entrapping them inside the printed structure in the same quantity as in the liquid.
A
A model cube structure (Fig. 2) served as a template to optimize the exposure time of the glass particle filled resin. Stereomicroscope images of the printed model were taken perpendicularly to the top of the printed samples and showed that a minimum exposure time of 1.5 s was necessary to fully fabricate the horizontal struts, which were as small as 100 µm. According to the different illumination tests, lower exposure times did not provide the necessary illumination to cure the smallest cube features, while higher printing times lead to an overexposure of the structure. Microscope images from the side of the model cube structure (Fig. 2) also showed that, at the selected optimal exposure time, the amount of photoabsorber present in the printing resin proved sufficient to limit the penetration depth of the printing light. No addition of further photoinitiator was necessary, as the exposure time selected was already suitable for the efficient printing of large structures.
IP T SC R U N A M ED PT CC E A Figure 1 Thermal analysis and XRD pattern of a printed and heat-treated sample.
Top view
Side view 1
Side view 2
2.0 s
SC R
Top view of cube structure at different exposure times (scale bars = 0.5 mm) 0.5 s 1.0 s 1.5 s
IP T
CAD model scaffold
U
Figure 2 Display of model cube structure (scale bars = 2 mm).
M
A
N
To ensure that the selected heating schedule would yield the desired phase assemblage in the bioceramic component, printed samples were crushed after sintering and analysed by XRD (Fig. 1). The samples yielded pure diopside (CaMgSi2O6) and wollastonite (CaSiO3) crystalline phases after sintering at 1100°C. No additional phase was detected in the heat-treated samples, showing a complete conversion into wollastonite-diopside biocompatible glass-ceramic.
PT
ED
The model cube structures, shown in Fig. 2, were also used to investigate the shrinkage of the wollastonite-diopside glass-ceramic printed samples during heating. A completely homogenous shrinkage in all directions was measured (linear shrinkage of 25.4 ± 1.6 %). TGA analysis (see Fig. 1) indicated that complete burn out of the photocured polymer occurred at 500°C. This means that during the polymer decomposition the glass particles kept the original shape and the green-network was strong enough to support the printed structure without collapsing. Furthermore, the efficient crystallization of the selected glass composition hindered any viscous collapse during the formation of the glass-ceramic, so that a complete shape retention was observed after firing.
A
CC E
After the optimization of printing parameters and heat treatment, unique complex structures, that could be manufactured solely by stereolithography, were printed. We show here samples possessing a Kelvin cell (a tetrakaidecahedron) structure, as an example of the architectural complexity that can be realized using stereolithography. Fig. 3 shows SEM images of the as printed and sintered samples. We can observe that all the features of the as printed sample, including the presence of discrete layers in z-direction, were completely transferred in the sintered sample. The images confirm that the uniform and isotropic shrinkage led to no deformation of the structures after the heat treatment.
IP T SC R U N A M
ED
Figure 3: SEM images of samples with a Kelvin cell architecture: as printed (A, B); sintered at 1100°C (C, D).
A
CC E
PT
The investigation of the fracture surface of a sintered glass-ceramic 3D printed sample (Fig. 4) showed that no internal nor surface microcracks could be found in the structure, indicating that the elimination of the photocurable polymer did not induce any macroscopic defects. Furthermore, it is worth observing that no delamination between the printed layers occurred, and that the fracture did not follow the path of the printed layers. This demonstrates that the interface between two layers was not the week point in the structure and that the individual layers were strongly connected. Finally, while individual layers could be observed on the surface of the printed and of the fired samples, no interfaces inside the structures were visible, suggesting that the viscous flow sintering and crystallization led to a sample possessing a homogeneous microstructure throughout its volume.
Figure 4. SEM image of the fracture surface of a sintered glass-ceramic printed sample.
IP T
The printed wollastonite-diopside glass-ceramic samples, after the heat treatment at 1100°C, possessed a bulk density of 0.47 g/cm3, true density of 3.12 g/cm3, and total porosity of 83 vol%. The apparent and true density were measured for printed glass-ceramic sample without base after sintering (Fig. 3C) and showed approximately the same value. Therefore, the struts have not closed porosity. Some amount of residual porosity could be observed on the surface of the sample as well as on the fracture surface. This could be attributed to the crystallization and transition into a glassceramic material, although the contribution of incomplete viscous flow and sintering to its presence should also be considered. However, it should be noted that this porosity in the struts, which is typical for a diopside and wollastonite glass-ceramic [18], can be useful for biological applications as a large surface area is beneficial for cell attachment and spreading [19, 20].
N
U
SC R
The cubic Kelvin cell structures (represented in Fig. 3) without base show a mechanical strength of 3.2 ± 0.5 MPa at a total porosity of 83 vol%. This demonstrates that the properties of the developed wollastonite-diopside glass-ceramic structures are within the standard reference range for the mechanical strength (2–12 MPa) [21] as well as for the porosity (75-95 vol%) [22] typical of human trabecular bone. The crushing strength is excellent, if we consider the classical model proposed by Gibson and Ashby [23]. For a bending-dominated structure, like the present one, the crushing strength (σc) depends on the bending strength (σbend) of the solid phase and on the relative density (ρrel) according to an exponential correlation: σc= σbend C ρrel3/2, where C is a dimensionless constant (being ≈0.2). From experimental value of σc and the low relative density (ρrel=0.17, for a total porosity of 83%) we can infer that the strength of the solid phase well exceeds 200 MPa.
A
4. Conclusions
CC E
PT
ED
M
Bioactive glass-ceramic samples possessing a highly complex architecture were fabricated using a photocurable polymer containing a high volume of glass particles, which crystallized into wollastonite and diopside upon heat-treatment. The glass particles were trapped in the cured polymer and highly porous structures were built up layer-by-layer via DLP, after optimization of the printing parameters. The heat-treatment in air, at 1100°C, caused the burn-out of the polymer and the sinter-crystallization of glass particles. While the material exhibited a residual porosity due to the formation of crystals, no delamination and microcrack formation could be observed in the structure. The samples showed a homogeneous shrinkage with complete shape retention. The bioactivity of wollastonite-diopside glassceramics, combined with the design freedom and high resolution of 3D stereolithography, as well as the good mechanical strength make these structures suitable candidates for bone tissue engineering applications.
Acknowledgments
A
J.S. gratefully acknowledges the CARIPARO Foundation, Padova, Italy, for her PhD scholarship. References
[1] S. Padilla, S. Sánchez-Salcedo, M. Vallet-Regí, Bioactive glass as precursor of designedarchitecture scaffolds for tissue engineering, Journal of Biomedical Materials Research Part A 81A(1) (2007) 224-232. [2] T.M. Freyman, I.V. Yannas, L.J. Gibson, Cellular materials as porous scaffolds for tissue engineering, Progress in Materials Science 46(3) (2001) 273-282.
A
CC E
PT
ED
M
A
N
U
SC R
IP T
[3] M.A. Sainz, P. Pena, S. Serena, A. Caballero, Influence of design on bioactivity of novel CaSiO3–CaMg(SiO3)2 bioceramics: In vitro simulated body fluid test and thermodynamic simulation, Acta Biomaterialia 6(7) (2010) 2797-2807. [4] B. Flautre, M. Descamps, C. Delecourt, M.C. Blary, P. Hardouin, Porous HA ceramic for bone replacement: Role of the pores and interconnections – experimental study in the rabbit, Journal of Materials Science: Materials in Medicine 12(8) (2001) 679-682. [5] A. Zocca, P. Colombo, C.M. Gomes, J. Günster, Additive Manufacturing of Ceramics: Issues, Potentialities, and Opportunities, Journal of the American Ceramic Society 98(7) (2015) 1983-2001. [6] N. Travitzky, A. Bonet, B. Dermeik, T. Fey, I. Filbert-Demut, L. Schlier, T. Schlordt, P. Greil, Additive Manufacturing of Ceramic-Based Materials, Advanced Engineering Materials 16(6) (2014) 729-754. [7] J.H. Jang, S. Wang, S.M. Pilgrim, W.A. Schulze, Preparation and characterization of barium titanate suspensions for stereolithography, Journal of the American Ceramic Society 83(7) (2000) 1804-1806. [8] J.W. Halloran, V. Tomeckova, S. Gentry, S. Das, P. Cilino, D. Yuan, R. Guo, A. Rudraraju, P. Shao, T. Wu, T.R. Alabi, W. Baker, D. Legdzina, D. Wolski, W.R. Zimbeck, D. Long, Photopolymerization of powder suspensions for shaping ceramics, Journal of the European Ceramic Society 31(14) (2011) 2613-2619. [9] Y. De Hazan, J. Heinecke, A. Weber, T. Graule, High solids loading ceramic colloidal dispersions in UV curable media via comb-polyelectrolyte surfactants, Journal of Colloid and Interface Science 337(1) (2009) 66-74. [10] T.-M. Chu, J. Halloran, S. Hollister, S. E. Feinberg, Hydroxyapatite Implants with Designed Internal Architecture, 2001. [11] S. Limpanuphap, B. Derby, Manufacture of biomaterials by a novel printing process, Journal of Materials Science: Materials in Medicine 13(12) (2002) 1163-1166. [12] K. Chopra, P. Mummery, B. Derby, J. Gough, Gel-cast glass-ceramic tissue scaffolds of controlled architecture produced via stereolithography of moulds, 2012. [13] V.I. Putlyaev, P.V. Evdokimov, T.V. Safronova, E.S. Klimashina, N.K. Orlov, Fabrication of osteoconductive Ca3-x M-2x (PO4)(2) (M = Na, K) calcium phosphate bioceramics by stereolithographic 3D printing, Inorg. Mater. 53(5) (2017) 529-535. [14] P. Tesavibul, R. Felzmann, S. Gruber, R. Liska, I. Thompson, A.R. Boccaccini, J. Stampfl, Processing of 45S5 Bioglass® by lithography-based additive manufacturing, Materials Letters 74 (2012) 81-84. [15] H. Elsayed, A. Rincón Romero, L. Ferroni, C. Gardin, B. Zavan, E. Bernardo, Bioactive Glass-Ceramic Scaffolds from Novel ‘Inorganic Gel Casting’ and Sinter-Crystallization, Materials 10(2) (2017) 171. [16] F. Baino, C. Vitale-Brovarone, Mechanical properties and reliability of glass–ceramic foam scaffolds for bone repair, Materials Letters 118 (2014) 27-30. [17] S. Zhang, N. Sha, Z. Zhao, Surface modification of α-Al2O3 with dicarboxylic acids for the preparation of UV-curable ceramic suspensions, 2016. [18] A. Karamanov, M. Pelino, Induced Crystallization Porosity and Properties of Sintereds Diopside and Wollastonite glass–ceramics, 2008. [19] O. Chan, M.J. Coathup, A. Nesbitt, C.Y. Ho, K.A. Hing, T. Buckland, C. Campion, G.W. Blunn, The effects of microporosity on osteoinduction of calcium phosphate bone graft substitute biomaterials, Acta Biomaterialia 8(7) (2012) 2788-2794. [20] A. Bignon, J. Chouteau, J. Chevalier, G. Fantozzi, J.P. Carret, P. Chavassieux, G. Boivin, M. Melin, D. Hartmann, Effect of Micro- and Macroporosity of bone substitutes on their mechanical properties and cellular response, 2004. [21] L.L. Hench, J. Wilson, An Introduction to Bioceramics, 2nd ed., Imperial College Press, 1993.
A
CC E
PT
ED
M
A
N
U
SC R
IP T
[22] M.A. Meyers, P.-Y. Chen, Biological Materials Science: Biological Materials, Bioinspired Materials, and Biomaterials, Cambridge University Press, England, 2014. [23] L.J. Gibson, M.F. Ashby, Cellular Solids: Structure and Properties, 2nd ed., Cambridge University Press, Cambridge, 1997.