Abstract
The crystallization behavior of racemic polylactide (equimolar PLLA/PDLA) blend spin-cast on thin film was investigated using FTIR and in-situ GISAXS/GIWAXS. The film thicknesses were controlled by coating with different solution concentration, which gives the film thicknesses equal to 630, 200, 65, and 16 nm. There is no crystalline found right after spin-coating, but crystalline peak would emerge during heating (from 40 °C to 250 °C at 3 °C/min) of thin films, which is same as cold crystallization process. It is discovered that racemic PLA blend will form into α and βc phase in 630 and 200 nm thick films, yet only form into βc phase in the films of 65 and 16 nm during cold-crystallization process. By practicing designed experiment, two possibilities are excluded to be responsible for the suppression in 65 and 16 nm thin film. For one hand, α phase is proved be able to form with only optically pure polylactide in 83 and 24 nm thin film. On the other hand, the spin-coating effect which may produce βc phase nuclei near substrate was erased by melt-quench process, but the same suppression of α phase in 65 and 16 nm result are still observed. From the observation of coherence length of α crystals that emerges in large size from the beginning of the GIWAXS peak becoming discernible, which may suggest that α phase have larger nuclei than βc phase. Moreover, when examine GIWAXS characteristic peaks carefully, the reflections of βc phase will emerge earlier than α phase, implying that formation of βc phase is earlier than α phase. The nanograin size of βc phase is ca. 10 nm initially, which is comparatively large in films under 100 nm. Therefore, earlier formation of βc phase will make formation of α phase is restricted by confined free space. The larger space which α phase required to form and the earlier formation of βc phase may be responsible for the suppression of α phase formation in 65 and 16 nm films. Besides, quantitative analysis for GIWAXS reveals that coherence length of crystalline reduced with decreasing film thickness. While the coherence length can be up to ca. 30 nm in the 630 nm thin film before melting, the coherence length in 16 nm thin film was only ca. 16 nm. Moreover, the orientation analysis from GIWAXS points out that the orientation in 630 nm film is worse than others, while 200, 65 and 16 nm thin film give apparent edge-on packing by thickness confinement. The orientation analysis also gives the idea that βc crystals may form in trigonal unit cell, and has its chain backbones lying on substrate. With model fitting analysis in GISAXS, the nanograins of βc phase can be well described with arrayed disks model. According to the fitting results, the radius of disk-like nanograins in 16 nm thin film is 2 times smaller than the one in 630 nm thin film, while the correlation between disks is two times better in 16 nm film than in 630 nm. Finally, by combining the GISAXS/GIWAXS results, a comparison crystal model for βc phase between 630 and 16 nm thin film was proposed.