The recent finding of large electrocaloric effect in several ferroelectric polymers creates unique opportunity for developing compact size solid state cooling cycles beyond the traditional mechanical vapor compression cycles. Here, we show that, by employing regeneration process with solid state regenerators, a chip scale Electrocaloric Oscillatory Refrigeration (ECOR) can be realized. A prototype ECOR is fabricated and characterized. More than 6 K temperature span is obtained near room temperature between the hot and cold sides of a 2 cm long device. Finite volume simulation validates the test results and shows the potential high performance of the ECOR.

We have implemented the Nosé-Hoover chain (NHC) method combined with an explicit reversible integrator into the MD module of AMBER (i.e., SANDER) to study the dynamics and the structure of biopolymers. We have implemented both constant temperature (NVT) and constant temperature and pressure (NTP) methods. We have studied the structure and dynamics of the antifreeze protein (AFP) in the gas phase and when solvated by water. Single and multiple chains of thermostats attached to the solute, solvent, and simulation box (in the case of constant-pressure simulations) were examined. The simulation results from constant energy, Berendsen constant temperature and pressure and Nosé-Hoover chain NVT and NTP methods indicate all these methods can evolve the system to equilibrium at a comparable rate. The NHC method controls temperature better over the other methods. In particular, separate thermostat chains can eliminate the cold solute-hot solvent problem. For the constant temperature and pressure simulations, the NHC method gives volume fluctuations that are in good agreement with the experimentally determined isothermal compressibility. The volume probability distribution of a system consisting of one free particle obtained using the NHC method agrees with the exact distribution very well. However, that obtained from the Berendsen method does not allow fluctuations to occur, and the average volume approaches the most probable value exponentially. The contribution of different types of interactions to the pressure is examined. We show that the potential energy of the bond angle and dihedral angle does not contribute to the virial calculated using an atom representation. We also note that the intermolecular interactions tend to deform the water molecules. Therefore, the H-O bond lengths of water molecules are longer and the H-H distance is shorter than the values set in the potential model if no bond constraints are present. This effect results in a large negative contribution to pressure due to the bond potential or constraint force.

We have studied the winter flounder antifreeze protein (AFP) and two of its mutants using molecular dynamics simulation techniques. The simulations were performed under four conditions: in the gas phase, solvated by water, adsorbed on the ice (2021) crystal plane in the gas phase and in aqueous solution. This study provided details of the ice-binding pattern of the winter flounder AFP. Simulation results indicated that the Asp, Asn, and Thr residues in the AFP are important in ice binding and that Asn and Thr as a group bind cooperatively to the ice surface. These ice-binding residues can be collected into four distinct ice-binding regions: Asp-1/Thr-2/Asp-5, Thr-13/Asn-16, Thr-24/Asn-27, and Thr-35/Arg-37. These four regions are 11 residues apart and the repeat distance between them matches the ice lattice constant along the (1102) direction. This match is crucial to ensure that all four groups can interact with the ice surface simultaneously, thereby, enhancing ice binding. These Asx (x = p or n)/Thr regions each form 5-6 hydrogen bonds with the ice surface: Asn forms about three hydrogen bonds with ice molecules located in the step region while Thr forms one to two hydrogen bonds with the ice molecules in the ridge of the (2021) crystal plane. Both the distance between Thr and Asn and the ordering of the two residues are crucial for effective ice binding. The proper sequence is necessary to generate a binding surface that is compatible with the ice surface topology, thus providing a perfect "host/guest" interaction that simultaneously satisfies both hydrogen bonding and van der Waals interactions. The results also show the relation among binding energy, the number of hydrogen bonds, and the activity. The activity is correlated to the binding energy, and in the case of the mutants we have studied the number of hydrogen bonds. The greater the number of the hydrogen bonds the greater the antifreeze activity. The roles van der Waals interactions and the hydrophobic effect play in ice binding are also highlighted. For the latter it is demonstrated that the surface of ice has a clathratelike structure which favors the partitioning of hydrophobic groups to the surface of ice. It is suggested that mutations that involve the deletion of hydrophobic residues (e.g., the Leu residues) will provide insight into the role the hydrophobic effect plays in partitioning these peptides to the surface of ice.

An integral-equation approach combined with molecular dynamics simulations based on the Girifalco spherical intermolecular potential has been used to predict the phase diagram for rigid C60 molecules. The boundary of the liquid-vapor coexistence region and the location of freezing and melting lines have been sketched out. The liquid phase is only observed in a very narrow temperature range compared with atomic systems (e.g., the rare gases). Unfortunately, the dense fluid is predicted to exist above -1800 K, which is sufficiently high that the C6o molecule may be unstable.

"Fail early and fail fast" is the current paradigm that the pharmaceutical industry has adopted widely. Removing non-drug-like compounds from the drug discovery lifecycle in the early stages can lead to tremendous savings of resources. Thus, fast screening methods are needed to profile the large collection of synthesized and virtual libraries involved in the early stage. Solubility is one of the filters that are applied extensively to ensure that the compounds are reasonably soluble so that synthesis of the compounds and assay studies of pharmacokinetics and toxicity are feasible. To address this need, we have developed a fast quantitative structure-property relationship (QSPR) model for the prediction of aqueous solubility (at 298 K, unbuffered solution) from the molecular structures. Multiple linear regressions and genetic algorithms were used to develop the models. The model was based on a set of diverse compounds including small organic molecules and drug and drug-like species. The predicted solubility for the training and test sets agrees well with the experimental values. The coefficient of determination is R(2) = 0.84 for the training set of 775 compounds and the RMS error = 0.87. This model was validated on four sets of compounds. The RMS error for the 1665 compounds from the four validation data sets (including compounds from the Physician's Desk References and Comprehensive Medicinal Chemistry databases) is 1 log unit and the unsigned error is 0.77. This model does not require 3-D structure generation which is rather time-consuming. Using 2-D structure as input, this model is able to compute solubility for 90 000-700 000 compounds/h on a SGI Origin 2000 workstation. This kind of fast calculation allows the model to be used in data mining and screening of large synthesized or virtual libraries.