Material model measurements and predictions for a random pore poly(epsilon-caprolactone) scaffold.

We investigated material models for a polymeric scaffold used for bone. The material was made by co-extruding poly(epsilon-caprolactone) (PCL), a biodegradable polyester, and poly(ethylene oxide) (PEO). The water soluble PEO was removed resulting in a porous scaffold. The stress-strain curve in compression was fit with a phenomenological model in hyperbolic form. This material model will be useful for designers for quasi-static analysis as it provides a simple form that can easily be used in finite element models. The ASTM D-1621 standard recommends using a secant modulus based on 10% strain. The resulting modulus has a smaller scatter in its value compared with the coefficients of the hyperbolic model, and it is therefore easier to compare differences in material processing and ensure quality of the scaffold. A prediction of the small-strain elastic modulus was constructed from images of the microstructure. Each pixel of the micrographs was represented with a brick finite element and assigned the Young's modulus of bulk PCL or a value of 0 for a pore. A compressive strain was imposed on the model and the resulting stresses were calculated. The elastic constants of the scaffold were then computed with Hooke's law for a linear-elastic isotropic material. The model was able to predict the small-strain elastic modulus measured in the experiments to within one standard deviation. Thus, by knowing the microstructure of the scaffold, its bulk properties can be predicted from the material properties of the constituents.

Constitutive models for a poly(e-caprolactone) scaffold.

We investigate material models for a porous, polymeric scaffold used for bone. The material was made by co-extruding poly(e-caprolactone) (PCL), a biodegradable polyester, and poly(ethylene oxide) (PEO). The water soluble PEO was removed resulting in a porous scaffold. The stress-strain curve in compression was fit with a phenomenological model in hyperbolic form. This material model will be useful for designers for quasi-static analysis as it provides a simple form that can easily be used in finite element models. The ASTM D-1621 standard recommends using a secant modulus based on 10% strain. The resulting modulus has a smaller scatter in its value compared to the coefficients of the hyperbolic model, and it is therefore easier to compare material processing differences and ensure quality of the scaffold. A third material model was constructed from images of the microstructure. Each pixel of the micrographs was represented with a brick finite element and assigned the Young's modulus of bulk PCL or a value of 0 for a pore. A compressive strain was imposed on the model and the resulting stresses were calculated. The elastic constants of the scaffold were then computed using Hooke's law for a linear-elastic isotropic material. The model was able to predict the small strain Young's modulus measured in the experiments to within one standard deviation. Thus, by knowing the microstructure of the scaffold, its bulk properties can be predicted from the material properties of the constituents.

Structure and mechanical properties of poly(D,L-lactic acid)/poly(epsilon -caprolactone) blends.

A series of blends of the biodegradable polymers poly(D,L-lactic acid) and poly( epsilon -caprolactone) were prepared by varying mass fraction across the range of compositions. Tensile testing was performed at room temperature using an extensometer and the elastic modulus was calculated for each blend. The blends were also tested to failure, and the strain-at-failure and yield stress recorded. While the blend has been shown to have a lower critical solution temperature, the mechanical properties were insensitive to the annealing conditions. Scanning electron microscopy was used to characterize the blend microstructure and poor adhesion was observed at the interface between blend components. Differential scanning calorimetry was performed but the results were somewhat variable, indicating this blend may have complex phase behavior that depends sensitively on the method of preparation. However, nuclear magnetic resonance data indicate the two components are phase separated. A percolation model is used to explain the observed mechanical data and the results are consistent with the predictions of the Kerner-Uemura-Takayangi model. The results of these experiments demonstrate the utility of polymer blending in tuning material properties.