Differences between Tethered Polyelectrolyte Chains on Bare Mica and Hydrophobically Modified Mica

This study investigates the structures of layers of amphiphilic diblock copolymers of poly(t-butyl styrene)-poly- (styrene sulfonate) (PtBS-PSS) adsorbed on both the bare mica surface (hydrophilic) and an octadecyltriethoxysilane (OTE)-modified mica surface (hydrophobic). When the surface is rendered hydrophobic, the nonsoluble block exhibits stronger interaction with the surface and higher adsorbed masses are achieved. Interaction forces between two such adsorbed layers on both substrates were measured using the surface forces apparatus. The effect of salt concentration (Cs) and molecular weight (N) on the height of the self-assembled layers (L0) was examined in each case. The resulting scaling relationship is in good agreement with predictions of the brush model, L0 ∞ N1.0 in the low-salt limit and L0N-1 ∞ (Cs/σ)-0.32 in the salted regime, when adsorption takes place onto the hydrophobized mica surface. For adsorption on the bare mica surface, L0N-0.7 ∞ Cs -0.17 agrees with the scaling prediction of the sparse tethering model. The results suggest that, on the hydrophilic bare mica surface, the adsorbed amount is not high enough to form a brush structure and only very little intermolecular stretching of the tethered chains occurs; in contrast, the presence of the hydrophobic OTE layer increases the tethering density such that the polyelectrolyte chains adopt a brush conformation.

An Extended Kernel for Generalized Multiple-Instance Learning

The multiple-instance learning (MIL) model has been successful in areas such as drug discovery and content-based image-retrieval. Recently, this model was generalized and a corresponding kernel was introduced to learn generalized MIL concepts with a support vector machine. While this kernel enjoyed empirical success, it has limitations in its representation. We extend this kernel by enriching its representation and empirically evaluate our new kernel on data from content-based image retrieval, biological sequence analysis, and drug discovery. We found that our new kernel generalized noticeably better than the old one in content-based image retrieval and biological sequence analysis and was slightly better or even with the old kernel in the other applications, showing that an SVM using this kernel does not overfit despite its richer representation.