scaffolds. The PLLA dioxane solutions in test tubes were cooled from the bottom surfaces at a constant rate by using liquid nitrogen to induce solid–liquid phase separation. The phase-separated samples were then dried under vacuum at -5 C for about 6 days to remove the solvent completely. Cylindrical samples were obtained and then trimmed to be a cylinder with 8-mm diameter and 12-mm length. Porosity (volume fraction of pores) of a scaffold was estimated from the densities of PLLA solid and the scaffold. Compression tests were then performed using a conventional mechanical testing machine at a loading rate of 1 mm/min. A fieldemission scanning electron microscope (FE-SEM) was also used to observe the microstructures of the porous samples and the deformation behavior of the porous samples at the critical point. FE-SEM micrographs of the cross-sections in the transverse direction are shown in Fig. 1. Homogeneous distribution of pores is observed, and the size of the pores tends to decrease with increase in the concentration of the solution. The porosities of these porous samples were estimated as 94.7, 90.0, and 88.5% for the concentrations of 3, 5, and 7 wt.%, respectively. FE-SEM micrograph of a crosssection in the vertical direction is shown in Fig. 2. For all the samples, this kind of porous structure was observed. These structures shown in Figs. 1 and 2 suggest that the crystals of the solvent grow in the vertical and transverse directions at a constant speed by the freezing process. Typical stress–strain curves under compression are shown in Fig. 3. The stress– strain relation of this porous structure of PLLA can be characterized by four different regions as shown in Fig. 4. The region A is the initial linear portion corresponding to linear elastic deformation of the porous structure. The end of the region A and the transition point to the region B is recognized as a critical point where localized failures of the wall structures of the pores take place. In the region B, stress