Substrate Roughness Speeds Up Segmental Dynamics of Thin Polymer Films

We show that the segmental mobility of thin films of poly(4-chlorostyrene) prepared under nonequilibrium conditions gets enhanced in the proximity of rough substrates. This trend is in contrast to existing treatments of roughness which conclude it is a source of slower dynamics, and to measurements of thin films of poly(2-vinylpiridine), whose dynamics is roughness invariant. Our experimental evidence indicates the faster interfacial dynamics originate from a reduction in interfacial density, due to the noncomplete filling of substrate asperities. Importantly, our results agree with the same scaling that describes the density dependence of bulk materials, correlating segmental mobility to a term exponential in the specific volume, and with empirical relations linking an increase in glass transition temperature to larger interfacial energy.

Figure 1

Structural relaxation of 20 nm thin films of P4ClS at $T=412\mathrm{K}$ (left panel), and P2VP at $T=403\mathrm{K}$ (right panel) deposited on Al substrates of different roughness. To ease reading only the contribution of segmental mobility is given in the main panels, while in the inset of the left panel we plotted the raw dielectric signal of films of P4ClS on the smoothest and the roughest substrate, solid lines through the experimental data are fits to the HN function [29]. In the inset of the right panel, power spectral density of the surface of Al substrates (see an example of 3D view in the bottom of the panel) with different $R_q$ values (blue squares 2 nm; green circles, 4 nm; red triangles, 9 nm) as a function of the wavelength k; dashed lines are fits to the $k$-correlation, $ABC$ function [13]. More details on fitting are provided in the Supplemental Material [25].

Figure 2

Upper panel, temperature dependence of the structural relaxation time for thin films of P2VP and P4ClS on substrates of different roughness; data of films ($h>300\mathrm{nm}$) sufficiently thick to exhibit bulk response are also given (open symbols). The value of $\tau$ for a 20 nm thin film of P4ClS ($R_q=5.1\mathrm{nm}$) annealed for more than 48 h at 433 K is indicated by a yellow star. Solid lines through experimental data are fits to the VFT equation. Lower panel, shift in dynamic glass transition for 20 nm thin films of P4ClS (pink triangles) and P2VP (green hexagons) as a function of the substrate roughness.

Figure 3

Upper panel, the volume of voids determined via the Abbot method (blue triangles) and the specific volume of the polymer layer (red square) obtained via the CFV model are given as a function of the substrate roughness. Lower panel, the inverse of the segmental relaxation time is plotted as a function of the substrate roughness. Based on these results, we provided a series of sketches of the interface of a polymer with substrates of different roughness. By increasing $R_q$ (left to right), the larger amplitude of the deviation from a flat substrate corresponds to a more pronounced reduction in density (pictured by less intense colors).