Gradient Extremal Methods

Gradient Norm Minimizations 14.5.9 Netvton-Raphson Methods 14.5.10 Gradient Extremal Methods Constrained Optimization Problems Locating the Global Minimum and Conformational Sampling 333 333 338 338 339 Appendix C First and Second Quantization Reference 411 411 412... 

Because (i) those parts of the free system PES that are actually explored by either classical trajectories or quantal wave packets describing a chemical reaction, are preferentially the lowest ones and (ii) modern quantum-chemical methods for calculating PES include geometry optimizations, generally gradient optimizations (for gradient extremal methods, see Refs.[64-66]), it is natural to retain for 3", at given 3, the values which minimize V(3). Mathematically, this is expressed as ... 

To develop a practical method for tracing gradient extremals we return to the restricted second-order model Eq. (3.10).17 In this model the Hessian is constant... 

The gradient of the PES (force) can in principle be calculated by finite difference methods. This is, however, extremely inefficient, requiring many evaluations of the wave function. Gradient methods in quantum chemistiy are fortunately now very advanced, and analytic gradients are available for a wide variety of ab initio methods [123-127]. Note that if the wave function depends on a set of parameters X], for example, the expansion coefficients of the basis functions used to build the orbitals in molecular orbital (MO) theory. 

The method has severe limitations for systems where gradients on near-atomic scale are important (as in the protein folding process or in bilayer membranes that contain only two molecules in a separated phase), but is extremely powerful for (co)polymer mixtures and solutions [147, 148, 149]. As an example Fig. 6 gives a snapshot in the process of self-organisation of a polypropylene oxide-ethylene oxide copolymer PL64 in aqueous solution on its way from a completely homogeneous initial distribution to a hexagonal structure. 

The method of preparation of a support material has a tremendous effect on its properties (11). For example, zeoHtes, which are highly stmctured aluminosihcates, are known to be extremely sensitive to the conditions employed both during and after crystallization (12). Also, when siUca—titania is precipitated by a coprecipitation method using ammonia, in which localized hydroxide ion gradients are estabUshed by the precipitation process itself, the product is much more acidic than when it is precipitated using urea, which suppHes hydroxide ion slowly and uniformly during precipitation (13). 

Instead of using repeated solution of a suitable eigenvalue equation to optimize the orbitals, as in conventional forms of SCF theory, we have found it more convenient to optimize by a gradient method based on direct evaluation of the ener functional (4), ortho normalization being restored after every parameter variation. Although many iterations are required, the energy evaluation is extremely rapid, the process is very stable, and any constraints on the parameters (e.g. due to spatial symmetry or choice of some type of localization) are very easily imposed. It is also a simple matter to optimize with respect to non-linear parameters such as orbital exponents. 

Fig. 3.4.3 (a) Two-dimensional k-space acquisition using SPI or SPRITE. The k-space data acquisition is indicated numerically. High magnetic field gradient amplitudes are applied at the extremities of k-space. (b) A generic two-dimensional centric scan SPRITE method. The... 

---