Homogeneous Bulk, Surface, and Edge Nucleation in Crystalline Nanodroplets

The birth of a crystal is initiated by a nucleus from which the crystal grows—a dust grain in a snowflake is a familiar example. These nuclei can be heterogeneous defects, like the dust grain, or homogeneous nuclei which are intrinsic to the material. Here we study homogeneous nucleation in nanoscale polymer droplets on a substrate which itself can be crystalline or amorphous. We observe a large difference in the nucleating ability of the substrate. Furthermore, the scaling dependence of nucleation on the size of the droplets proves that the birth of the crystalline state can be directed to originate predominantly within the bulk, at the substrate surface, or at the droplets’ edge, depending on how we tune the substrate.

Figure 1

(a) Schematic diagram of the sample geometry. $i$-PS film is prepared in the crystalline ($ci$-PS) or amorphous ($ai$-PS) state. Droplets of PEO are formed after the as-prepared sample (top) is allowed to dewet (bottom). (b) Optical microscopy image (width $400\mu \mathrm{m}$) of a typical sample showing the dewetted droplets. Crystalline droplets appear bright.

Figure 2

(a) Typical correlation plot of the temperature at which a drop crystallizes in subsequent cooling runs. The PEO droplets on an amorphous $i$-PS substrate ranged in base radius from 3.5 to $7.5\mu \mathrm{m}$, and were cooled from the melt at $1°\mathrm{C}/\min$. (b) The fraction of crystallized droplets upon cooling for three types of substrates.

Figure 3

(a)–(c) Logarithm of the amorphous fraction of droplets as a function of time for isothermal crystallization. The data are binned according to base area of droplets supported on the $ai$-PS substrate, $ci$-PS(185) substrate, and $ci$-PS(175) substrates, respectively. (d) Logarithmic plot of the time constant for homogeneous nucleation as a function of the radius of the base of the droplets obtained from (a)–(c). The differences in scaling indicate that bulk, surface, and line nucleation are observed. Since each sample type is crystallized at a different temperature, there is no significance to the crossing of the lines.

Figure 4

AFM images (error signal) of crystalline PEO droplets on $i$-PS substrates (width $\sim 8\mu \mathrm{m}$). (a) Droplet on $ai$-PS substrate, with the nucleus (arrow) visible at the center. (b),(c) Droplets on $ci$-PS(175) substrate. In both cases the nucleus originates at the contact line.