Expectation values linear response

In the previous section we described the result of turning on a perturbation on the wave functions (eigenvectors) of the unperturbed Hamilton operator with nondegenerate spectrum in the lowest order when this effect takes place. In quantum mechanics the wave function is an intermediate tool, not an observable quantity. The general requirement of the theory is, however, to represent the interrelations between the observables. For this we give here the formulae describing the effect of a perturbation upon an observable. Let us assume that in one of its unperturbed states the system is characterized by the expectation value of an observable A  

Turning on the perturbation A W produces the correction to the wave functions (eigenvectors) of the system described by eq. (1.62). Inserting it into the definition of the expectation value of A yields  

The coefficient at A describes the linear response of the quantity A to the perturbation W. It can be given a rather more symmetric form. Indeed the amplitude of the j-th unperturbed state in the correction to the fc-th state is proportional to some skew Hermitian operator (the perturbation matrix W is Hermitian, but the denominator changes its sign when the order of the subscripts changes). With this notion and assuming that Wkk = 0 (see above) we can remove the restriction in the summation and write  

It is of interest to learn more about the operator K which plays that remarkable r61e. It is defined by the perturbation operator W and the unperturbed Hamiltonian H 0). As a function of W it is obviously linear in the sense that if W = U  

With its use the response of the quantity A in the k-th state of the operator H° J to the perturbation given by the operator W is conveniently written as  

---

Only static and dynamic molecular properties involving electric dipole and quadrupole operators will be discussed below. However, electric properties related to higher-order electric multipole operators can also be determined in a similar manner to the properties described here, in terms of expectation values, linear and nonlinear response functions. Nevertheless, it should be kept in mind that although the same formalism is applied in the calculation of response functions involving octupole, hexadecapole, and higher moments, in practice it may... 

The last method used in this study is CCSD linear response theory [37]. The frequency-dependent polarizabilities are again identified from the time evolution of the corresponding moments. However, in CCSD response theory the moments are calculated as transition expectation values between the coupled cluster state l cc(O) and a dual state... 

Perturbative estimate of ESVs with respect to noncorrelated bare Hamiltonian. The specificity of each bond and molecule in the approach based on the SLG expressions for the wave function is taken into account perturbatively by using the linear response approximation [25]. We need perturbative estimates of the expectation values of the pseudospin operators which, in their turn, give values of the density matrix elements according to eq. (3.5). According to the general theory (Section 1.3.3.2) the linear response 5(A) of an expectation value of the operator A to the time independent perturbation AB of the Hamiltonian (A is the parameter characterizing the intensity of the perturbation) has the form ... 

The first term on the right-hand side is the time-independent expectation value, whereas the second, third, and fourth terms describe the linear, quadratic, and cubic response to the perturbation, respectively. The Fourier transformed representations are given by... 

Two level factorial designs are primarily useful for exploratory purposes and calibration designs have special uses in areas such as multivariate calibration where we often expect an independent linear response from each component in a mixture. It is often important, though, to provide a more detailed model of a system. There are two prime reasons. The first is for optimisation - to find the conditions that result in a maximum or minimum as appropriate. An example is when improving die yield of synthetic reaction, or a chromatographic resolution. The second is to produce a detailed quantitative model to predict mathematically how a response relates to die values of various factors. An example may be how the near-infrared spectrum of a manufactured product relates to the nature of the material and processing employed in manufacturing. 

Just as in simple linear regression it is assumed that the random errors are independent of mean zero and variance. Under these assumptions, the expected value of the response for the r set of conditions for the regressors is ... 

There are, however, many pieces of evidence that the cysteine link only causes small perturbations in the electronic structure and energetics of tyrosine. Electrochemical experiments by Whittaker et al [31] showed that the p fiTa of o-methylthiocresol was only 0.7 pH units lower than for cresol (9.5 vs. 10.2). Babcock and co-workers have shown, based on EPR and ENDOR experiments on both apo-enzyme and model alkylthio-substituted phenoxyl radicals, that the sulfur cross-link only induces small perturbation in the spin distribution of the tyrosyl radical [32]. No big shift in the g-tensors between unsubstituted and methylthio-substituted radicals was observed. Since this kind of shift is expected when heavy elements carry some of the spin in organic radicals, the conclusion was that the sulfur center possesses only a small part of the unpaired spin. We have conducted ab initio multiconfigurational linear response g-value calculations of unsubstituted and sulfur-substituted phenoxyl radicals and shown that the shift in g-tensor is as small as 0.0008 in the gxx-component (2.0087 vs. 2.0079 in t). The other components were virtually unchanged, thus confirming the experimental results [33]. 

The first term is the expectation value and from the fact that the density change is the static linear response function... 

At the moment of writing very few implementations of the theory of molecular properties at the 4-component relativistic molecular level, beyond expectation values at the closed-shell Hartree-Fock level, have been reported. The first implementation of the linear response function at the RPA level in a molecular code appears to be to MO-based module reported by Visscher et al. [97]. Quiney and co-workers [98] have reported the calculation of second-order properties at the uncoupled Hartree-Fock level (see section 5.3 for terminology). Saue and Jensen [99] have reported an AO-driven implementation of the linear response function at the RPA level and this work has been extended to quadratic response functions by Norman and Jensen [100]. Linear response functions at the DFT/LDA-level have been reported by Saue and Helgaker [101]. In this section we will review the calculation of linear and quadratic response functions at the closed-shell 4-component relativistic Hartree-Fock level. We will follow the approach of Saue and Jensen [99] where the reader is referred for further details. 

The calculation of the dynamic polarizabilities by using the TDHF equar tions described above is equivalent to the Random Phase Approximation (RPA) within the linear response theory, LRT. > ° The subject of the linear response theory is the first-order change in the expectation value, 5 (A), of a property A, as response of the system to a perturbing field For... 

The main effects are twice the 61 and 62 coefficient values in the regression, because while 61 and 62 represent changes in the response caused by unit variations in xi and X2, in the effects the changes correspond to two units, from Xj = -1 to +1. The interaction effect is much smaller than the main effects. This is to be expected, since the response surface is adequately represented by a linear model. 

---