Interface Structure of Diblock Copolymer Brushes and Surface Dynamics of Homopolymer Brushes and Bilayers of Untethered Chains on Brushes

The main objective of the work presented in this dissertation is to understand the internal structure and surface dynamics of diblock copolymer brushes (DCBs). DCBs have garnered enormous interest in recent years due to their stimuli-responsive behavior. The characteristics of a surface of a DCB can be changed reversibly from those of one type to those of another by using a selective solvent for one block. There has not been any experimental study to prove either the dry state structure or the surface rearranged structure or the mechanism of that rearrangement. Understanding the dynamics of the brush/air interface and internal interfaces is also important to understanding the switching mechanism, and the dynamics at the air surface have further implications for wetting, friction and adhesion.

Using neutron reflectivity (NR), we have resolved the internal structure of DCBs of polystyrene-b-poly(methyl acrylate) (PS-b-PMA) and PMA-b-PS. The internal brush structure depends strongly on the block sequence and the value of χ^N (χ: Flory-Huggins interaction parameter, N: length of the chains). For dPS-b-PMA films, when χ^N ≤ 23, a model of two layers with an interfacial region of finite width provides a good description of the data. When χ^N is higher, a third layer must be included in the model. A third layer is consistent with the presence of a lateral ordering in the center of the brush, as evidenced by correlation peaks in grazing incidence small angle scattering (GISAXS) data from these samples. For PMA-b-dPS samples,when χ^N > 23, the change in reflectivity due to a 3-D structure in the middle of the film may be modeled with a dramatic increase of interface width.

In-situ NR measurements of DCBs in selective solvents show that the structure changes in a way to minimize the unfavorable interaction between solvent and the solvophobic block. How the chains in a DCB minimize this unfavorable interaction depends also on the solvent quality, the ratio of chain lengths of the polymer blocks, block sequence and details of the brush composition determined by nonidealities in the synthesis DCBs. A switching of the surface character is not easily obtained unless optimum conditions are provided. More work with DCBs of various block lengths should be pursued. When DCBs are swollen with a solvent good for both blocks, they behave in a manner similar to that for homopolymer brushes, with scattering length density (SLD) prole having, overall, a parabolic shape.

The dynamics of surface height fluctuations of molten homopolymer brushes have been investigated using X-ray photon correlation spectroscopy (XPCS). Within the range of time (0.2 -1100 s) and length scale (0.6-3 µm) studied, no relaxation of fluctuations of the brush surfaces are detectable. This is true for PS brushes of thicknesses of 9 - 101 nm and high grafting density (>0.5 chains/nm²) at temperatures up to 130°C above bulk T_g. Even reduction of the grafting density from 0.6 to 0.1 chains/nm² does not bring the relaxation into the measurement window.

Tethering not only changes the surface dynamics of grafted chains but also the surface dynamics of 2.2k dPS chains spun cast on top of a PS brush. The relaxation of surface height fluctuations on a melt of a certain quantity of PS chains spun cast on top of a PS brush becomes slower than the relaxation of surface height fluctuations on a melt of the same quantity of PS chains placed on top of a bare silicon surface. As the thickness of the brush layer increases (h_b), the relaxation time (τ) of fluctuations on top of the “bilayer” becomes larger. A model of single layer of viscous fluid on a rigid substrate with a non-slip boundary condition does a good job of capturing the data. Calculated thicknesses from the model that are “effective” for the observed surface dynamics are decreases with increasing h_b.

NR measurements reveal the actual structure of the bilayer lms. The structure model consists of a bottom layer of one SLD, a top layer of a second SLD, and an interface between the two layers is found to be reasonable. The “top” layer thicknesses observed with NR are the same as the “effective” thicknesses from the single viscous layer analysis of the XPCS data within the uncertainties. As the grafting density of the brush is reduced, the surface dynamics of the layer of spun cast chains on top of the brush become slower. This is due to increasing of free chains into the brush and a decrease in the “effective” thickness of the layer of free chains on top.

Increasing the molecular weight of the free chains, M_n, decreases the interface width. Interface width is a decreasing function of M_n and decays exponentially with increase of M_n.