Fabrication of a Porous and Formable Ceramic Composite Bone Tissue Scaffold at Ambient Temperature

Abstract

Broken or diseased bone tissue requires a multitude of repair strategies ranging from autografts, to allografts, to synthetic bone grafts. Herein we describe a fabrication approach to create all-ceramic resorbable porous scaffolds for bone tissue repair in the absence of traditional consolidation techniques for ceramics—a technique that has potential for in situ use in operating rooms, or in field hospitals. The room temperature and pressure process utilizes a reaction with a liquid ceramic precursor to form a silicate-glass binder phase to consolidate bioactive glass frit particles and make a formable paste. This formable paste is designed to be applied directly to a bone defect, hardening in situ to become a rigid scaffold and acting as a bone ‘spackling’ paste. Characterization of the composite scaffolds is evaluated with respect to design specifications required in biomedical implant materials, namely: formability, geometric stability, porosity, and load-bearing capacity. Of the fabricated scaffolds, the composite scaffold with 7.4 vol pct sodium silicate binder was found to have the most dispersed, well distributed, and largest amount of open porosity (44 pct), average compressive strength of 1.3 MPa (with the least amount of variability), surface area to volume ratio of 33 mm⁻¹ (slightly greater than trabecular bone), and excellent geometric stability. Results of this study indicate that the developed fabrication method should be further explored for synthetic bone graft materials.

Appendix: Pycnometry Data Analysis

Assuming that the minimum volume that a composite scaffold could occupy would equal the combined volume of the 45S5 frit (V_BAG) and solid sodium silicate (V_WG,s), it then follows that the pycnometry volume measurement of a specimen with 100 pct open porosity would equal this ‘minimum volume’ value (V_min). If the measured volume of a specimen is greater than this minimum volume, it is assumed that the additional volume is exclusively comprised of closed pores.

The minimum volume values were calculated as follows:

$$ \begin{array}{*{20}c} {V_{\hbox{min} } = V_{\text{BAG}} + V_{\text{WG,s}} } \\ \end{array} $$

(A1)

$$ \begin{array}{*{20}c} {V_{\text{BAG}} = \frac{{\left( {m_{{{\text{tested}}\,{\text{specimen}}}} } \right) \cdot \left( {{\text{wt\, pct}} {\text{BAG}}_{\text{set}} } \right)}}{{\rho_{\text{BAG}} }}} \\ \end{array} $$

(A2)

$$ \begin{array}{*{20}c} {V_{\text{WG, s}} = \frac{{\left( {m_{{{\text{tested}}\,{\text{specimen}}}} } \right) \cdot \left( {{\text{wt \,pct}} WG_{\text{set}} } \right)}}{{\rho_{\text{WG, s}} }}} \\ \end{array} $$

(A3)

$$ \begin{array}{*{20}c} {{\text{wt pct BAG}}_{\text{set}} = 1 - {\text{wt pct WG}}} \\ \end{array}_{\text{set}} $$

(A4)

$$ \begin{array}{*{20}c} {{\text{wt pct WG}}_{\text{set}} = \frac{{m_{\text{WG,s}} }}{{m_{WG,s} + m_{\text{BAG}} }}} \\ \end{array} $$

(A5)

$$ \begin{array}{*{20}c} {m_{\text{WG, s}} = \left( {m_{\text{WG, l}} } \right) \cdot \left( {{\text{mass}}\,{\text{loss}}\,{\text{factor}}} \right)} \\ \end{array} $$

(A6)

$$ \begin{array}{*{20}c} {{\text{mass}}\,{\text{loss}}\,{\text{factor}} = \left( {1 - \frac{{m_{\text{WG,l}} - m_{\text{WG,s}} }}{{m_{\text{WG,l}} }}} \right)} \\ \end{array} $$

(A7)

Bioactive glass density (ρ_BAG), set sodium silicate density (ρ_WG,s), and the mass loss factor were determined empirically via gas pycnometry.

The specimen mass (m_{tested specimen}) used in this calculation was the mass measured at the time of pycnometry analysis, as opposed to the combined weight of the measured feedstock materials. This convention accounted for the effect of the material loss—during the manufacturing process, residual paste adheres to the weigh boat, the bulb of the pipette used for mixing, and to gloves used to handle the specimen. Additionally, the friable nature of the set specimen can lead to loss of solid material during handling after setting.

Assumptions:

• (1) The material lost during manufacturing and handling was assumed to be lost proportionally; i.e., any lost material has the same composition as the bulk of the specimen. Realistically, the unset waterglass wets to the weigh boat, and it is likely that this wetting leads to a disproportionate loss of waterglass. It follows that compositions with lower amounts of waterglass may be more affected by this disproportionate loss.

• (2) At the time of analysis, the sodium silicate was assumed to be fully set, and assumed to not densify or shrink further. The time elapsed between manufacturing and analysis varied between specimen groups, and may have had an effect on the degree to which to sodium silicate densified. To mitigate this variability, kinetics studies examining the rate of mass loss during setting of sodium silicate were used to confirm the stability of the specimens prior to gas pycnometry analysis. As shown in Figure A1 in the Appendix, sodium silicate in air reaches half of its maximum normalized mass loss (i.e., setting reaction of binder is 50 pct complete) within 13.76 hours (parameter B in the curve fit). All pycnometry samples were allowed to set for a minimum of 288 hours before testing; as indicated on Figure A1 in the Appendix by the vertical dashed line, this setting time allowed the binder phase to reach the stable asymptotic region of the kinetic curve (95 pct level of maximum normalized mass loss).

  Fig. A1

  

  Kinetics data and curve fit for the mass loss during setting of bulk aqueous sodium silicate in air at 294 K (21 °C). The curve fit (red line) is based on the Michaelis–Menten equation describing a reaction as a function of concentration of reactants which decreases in time (as the reactants are consumed—much like is occurring in the setting of sodium silicate). The horizontal dashed line represents the asymptotic value of the mass loss (parameter A) and the vertical dashed line represents the setting time used for pycnometry specimens in air. Parameter B in the curve fit represents the time to reach half the maximal mass loss. Bulk solution kinetics were used as a conservative control due to the lowest SA/V ratio as compared to the total internal area of the liquid spread between frit particles in the composite

• (3) The measured densities of solid sodium silicate and 45S5 bioactive glass are assumed to be representative of the densities of these constituents in the composite specimens. In actuality, the geometry of the setting waterglass was different (i.e., droplets on parafilm vs. contained within a composite scaffold), which may have affected the measured density. Additionally, the total set time for the sodium silicate specimen was different than the total set time for each composite specimen.

• (4) It is assumed that no reaction occurs between the two constituents that would alter the density of either constituent.

• (5) The sodium silicate-only specimen that was analyzed was assumed to have no closed pores; closed pores would alter the empirically determined ρ_WG,s value.