Symmetrical experimental domain

Response surface designs can be divided into symmetrical and asymmetrical designs (7). The first type examines the factors in a symmetrical experimental domain, while the second can be chosen when an asymmetrical experimental domain is to be examined. 

As for the symmetrical designs and in agreement with the philosophy of experimental designs, the experimental domain is mapped as well as possible. This explains why, except for a central point, often all experiments of the D-optimal design are situated toward the boundaries of the experimental domain (Figure 2.10d). During method optimization, D-optimal designs with a symmetrical experimental domain were applied in References 19,60, and 95, and with an asymmetrical experimental domain in Reference 92. 

The way to ensure a clean extraction of an experimental reference signal is thus to zero-fill the experimental free induction decay s it) once before Fourier transformation, zero completely the imaginary part of the resultant spectmm, and zero all but the reference region to wr of the real part [7], Inverse Fourier transformation then gives a symmetric time-domain signal, the first half of which is the required experimental reference signal Sr t) ... 

The experimental domain can be symmetric or irregular. In the first situation the classical symmetrical designs are applied, while in the second non-symmetrical designs are constructed. 

This observation is expected from theory, as the observed thickness distributions are exactly the functions by which one-dimensional short-range order is theoretically described in early literature models (Zernike and Prins [116] J. J. Hermans [128]). From the transformed experimental data we can determine, whether the principal thickness distributions are symmetrical or asymmetrical, whether they should be modeled by Gaussians, gamma distributions, truncated exponentials, or other analytical functions. Finally only a model that describes the arrangement of domains is missing - i.e., how the higher thickness distributions are computed from two principal thickness distributions (cf. Sect. 8.7). Experimental data are fitted by means of such models. Unsuitable models are sorted out by insufficient quality of the fit. Fit quality is assessed by means of the tools of nonlinear regression (Chap. 11). 

Fig. 18 Phase space of PI-fc-PS-fc-PEO in vicinity of ODT. Filled and open circles-. ordered and disordered states, respectively, within experimental temperature range 100 < T/° C< 225. Outlined areas compositions with two- and three-domain lamellae (identified by sketches) shaded regions three network phases, core-shell double gyroid (Q230), orthorhombic (O70), and alternating gyroid (Q214). Overlap of latter two phase boundaries indicates high- and low-temperature occurrence, respectively, of each phase. Dashed line condition tfin = 0peo associated with symmetric PI-fc-PS-fc-PEO molecules. From [75]. Copyright 2004 American Chemical Society...

The exponent m cannot be regarded as a fitting parameter but depends on the symmetry of the system. In most cases, m = 3/2 [16, 140, 158, 166, 167, 174, 175], but m = 2 for highly symmetric systems, such as aligned Stoner-Wohlfarth particles. In particular, the m = 3/2 law is realized for misaligned Stoner-Wohlfarth particles and for most domain-wall pinning mechanisms [5], Experimental values of m tend to vary between 1.5 to 2 [136, 158]. Linear laws, where m = 1, are sometimes used in simplified models, but so far it hasn t been possible to derive them from physically reasonable energy landscapes [5, 16, 176]. The same is true for dependences such as /H- l/H0 [177], where series expansion yields an m = 1 power law. 

A mean field theory has recently been developed to describe polymer blend confined in a thin film (Sect. 3.2.1). This theory includes both surface fields exerted by two external interfaces bounding thin film. A clear picture of this situation is obtained within a Cahn plot, topologically equivalent to the profile s phase portrait d( >/dz vs < >. It predicts two equilibrium morphologies for blends with separated coexisting phases a bilayer structure for antisymmetric surfaces (each attracting different blend component, Fig. 32) and two-dimensional domains for symmetric surfaces (Fig. 31), both observed [94,114,115,117] experimentally. Four finite size effects are predicted by the theory and observed in pioneer experiments [92,121,130,172,220] (see Sect. 3.2.2) focused on (i) surface segregation (ii) the shape of an intrinsic bilayer profile (iii) coexistence conditions (iv) interfacial width. The size effects (i)-(iii) are closely related, while (i) and (ii) are expected to occur for film thickness D smaller than 6-10 times the value of the intrinsic (mean field) interfacial width w. This cross-over D/w ratio is an approximate evaluation, as the exact value depends strongly on the... 

A graphical approach was also used by Millet and Pons to analyse anisotropy of rotational diffusion in proteins. The values of Z)j and DJD compatible with R IRi ratios are presented as a contour plot. The intersection of the contour plots for different residues provides the values of anisotropy parameters compatible with experimental data. The obtained parameters can be used as starting values for further optimisation. The method is apphcable to axially symmetric rotation. A combination of approximate and exact methods was used by Ghose et al. to reduce the computational time of the determination of rotational diffusion tensor from NMR relaxation data. The initial values of the tensor components and its orientation are evaluated from the approximate solution, which substantially reduces the search space during the exact calculations. The method was applied for the estimation of relative domain orientation of a dual domain protein. 

---