Graphene oxide was synthesized from natural graphite by a modified Hummers method. 1 5 g graphite and 2.5 g sodium nitrate (NaNO 3 ) were mixed with 160 mL of concentrated sulfuric acid (H 2 SO 4 , 95%, Sigma-Aldrich) in a 500 mL flask.

The influence of solvent viscosity on the surface and internal structural dynamics of the protein myoglobin is studied using ultrafast infrared vibrational echo measurements of the pure dephasing of the A 1 CO stretching mode of myoglobin-CO (Mb-CO). The dephasing reflects protein structural fluctuations as sensed by the CO ligand bound at the protein's active site. Measurements made as a function of solvent viscosity at 295 K show that the pure dephasing has a marked dependence on viscosity. In addition, the pure dephasing of Mb-CO in the solvents trehalose and 50:50 ethylene glycol:water are compared as a function of temperature T (10-295 K). The pure dephasing data in the two solvents have identical T 1.3 temperature dependences at low temperatures, where both solvents are glassy solids. At higher temperatures, the Mb-CO pure dephasing has a much steeper temperature dependence in ethylene glycol:water, which is a liquid, than in trehalose, which is a glass at all temperatures studied. The steep temperature dependence in liquid ethylene glycol: water is described as a combination of a viscosity-dependent component and a temperature-dependent component. The viscosity-dependent data are analyzed using a theory that connects the fluctuations of the protein surface to the solvent's viscoelastic response. When the solvent's viscosity is lowered, the increased rate of fluctuation of the protein's surface allows more rapid internal protein dynamics, which result in more rapid dephasing. Good agreement is obtained for physically reasonable parameters. The experimental echo decay times are proportional to the cube root of the solvent viscosity η 1/3 . This proportionality is characteristic of protein structural fluctuations that give rise to CO frequency fluctuations that are in the spectral diffusion regime (relatively slow evolution).

In recent years, graph/hypergraph-based deep learning methods have attracted much attention from researchers. These deep learning methods take graph/hypergraph structure as prior knowledge in the model. However, hidden and important relations are not directly represented in the inherent structure. To tackle this issue, we propose a dynamic hypergraph neural networks framework (DHGNN), which is composed of the stacked layers of two modules: dynamic hypergraph construction (DHG) and hypergrpah convolution (HGC). Considering initially constructed hypergraph is probably not a suitable representation for data, the DHG module dynamically updates hypergraph structure on each layer. Then hypergraph convolution is introduced to encode high-order data relations in a hypergraph structure. The HGC module includes two phases: vertex convolution and hyperedge convolution, which are designed to aggregate feature among vertices and hyperedges, respectively. We have evaluated our method on standard datasets, the Cora citation network and Microblog dataset. Our method outperforms state-of-the-art methods. More experiments are conducted to demonstrate the effectiveness and robustness of our method to diverse data distributions.