Reference force

Gain ratio 17 r can be calculated at a reference force ratio, such as xopt, which is a natural steady-state force ratio of oxidative phosphorylation. This is seen as a result of the adaptation of oxidative phosphorylation to various metabolic conditions and also as a result of the thermodynamic buffering of reactions catalyzed by enzymes. The experimentally observed linearity of several energy converters operating far from equilibrium may be due to enzymatic feedback regulations with an evolutionary drive towards higher efficiency. 

F(r ri) is the Foiuier-space contribution of the force between two unit charges at positions ri and ri -i- r as calculated by the P M method (note that due to broken rotational and translational symmetry this does in fact depend on the coordinates of both particles), and R(r) is the corresponding exact reference force (whose Fourier transform is just Eq. 23). The inner integral over r scans all particle separations, whereas the outer integral over ri averages over all possible locations of the first particle within a mesh cell. Obviously, up to a factor L this expression is just the mean-square error in the force for two unit charges, in other words, the quantity x from Eq. 28. This provides a link between the rms error of an N particle system and the error Q from Hockney and Eastwood. Using Eq. 32 one obtains... 

Admittedly, Eq. 35 looks rather complicated. Still, in combination with Eq. 34 it gives the rms force error of the P M method or, more precisely, of its Fourier-space contribution. After all, the computation of Qopt and that of Gopt are quite similar. It should be emphasized that the formula Eq. 35 for the optimal Q value, just like the one for the optimal influence function in Eq. 22, is of a very general nature. It also works for different charge assignment functions, reference forces, or any differentiation scheme which can be expressed by an operator D(k). 

The components of the force between an anion and a cation inside the short-range domain [Iyc = 2 nm) are shown in Figure 11 as a function of the interionic separation. The two ions are placed in a 500-mM KCl solution, with no external bias. As expected, the reference force and mesh force have the same amplitude and therefore will cancel within the short-range domain. 

The force calculated n times (fast force) is called the reference force. A sample code of an MTS integrator is given in Algorithm 2. 

Two modifications of the standard RESPA method exist, depending on the application of the extended operator exp(iL vr/0- Tlie first variant of RESPA is useful when the evolution prescribed by the operator exp(iLNHt) is slow compared to the time scale associated with the reference force. It is formed by writing... 

Above equation is an over-determined system of linear equations if MParrinello molecular dynamics (CP-MD) or any other quantum chemical or classical MD simulations. L is the number of conformations, by sampling more a large number of L of atomic configurations can be obtained and therefore, the system of equations can be over-determined. Algorithms like QR to solve such over-determined equation in a least squares sense are used. 

This method was initially tested with water simulations where FM was done with reference force from ab-initio CP-MD calculations. Further this method was applied to liquid state systems by matching the force from classical all atomistic MD to CG MD. The same method was also tested for biomolecules such as dimyristoylphosphatidylcholine (DMPC) lipid bilayer and ionic liquids. 

Longuet-Higgins and Salem [1,8] proposed a molecular force field for localized a- electrons and delocalized 77 electrons. Their scheme has become attractive for conjugated polymers [25,26], now in conjunction with semiempirical rather than simple Hiickel calculations. Warshel and Karplus [27] and Hemley et al. [28] combine all-electron calculations of geometry with PPP models for 77 electrons (see also Ref. 21). We will focus exclusively on delocalization or 77-electron contributions using linear response (LR) theory, as illustrated by the AM formalism. LR has been widely applied to vibrational spectra of charge transfer and ion-radical organic crystals [29-32]. The idea is to model shifts due to delocalization relative to some localized reference. The concomitant problem of localized or other contributions is the choice of a reference force field discussed in Section II.B. The solid-state perspective of LR theory is quite compatible with 77-electron or other models based on frontier orbitals. 

Our goal is to model quantitatively 7r-electronic contributions to both vibrational and electronic spectra. The general e-ph analysis introduced in Section II combines the microscopic AM formalism [18,19] with the spectroscopic ECC model [22]. The reference force field F for PA provides an experimental identiHcation of delocalization effects. Transferable e-ph coupling constants are presented in Section III for polyenes and isotopes of trans- and a s-PA. The polymer force field in internal coordinates directly shows greater delocalization in t-PA, while coupling to C—C—C bends illustrates V(/ ) participation and different coupling constants a(/ a) and a(Jis) in Eq. (3) support an exponential r(/ ). NLO spectra of PDA crystals and films are presented in Section IV, with multiphoton resonances related to excited states of PPP models and vibronic contributions included in the Condon approximation. Linear and electroabsorption (EA) spectra of PDA crystals provide an experimental separation of vibrational and electronic contributions, and the full tt-tt spectrum is needed to model EA. We turn in Section V to correlated descriptions of electronic excitations, with particular attention to theoretical and experimental evidence for one- and two-photon thresholds of centrosymmetric backbones. The final section comments on parameters for conjugated polymers, extensions, and open questions. 

Once the reference force field F is properly derived, diagonalization of the GF° matrix [see Eq. (8)] yields the reference frequencies normal coordinates Q, . We use capital Q s to distinguish normal coordinates from the generic nuclear coordinates q in Eq. (9). This is a partial solution, since we still have to account for linear e-ph coupling. The expansion in Eq. (9) is conveniently carried out on the basis of the reference normal coordinates. The simplest case arises when, at least for a given symmetry subspace, only one electronic operator, 0, is coupled to the phonons. 

Fig. 6.4 Relative x vs. w curves, Eq. (14), for three Ug modes of rra/i5-polyacetylene coupled to 7r-electron fluctuations. The X = 0 0 and 1.0 values are fixed by the reference force field and the coupling constants gi. The IR frequencies of doped and photoexcited samples fall on the curves at larger , while the Raman frequencies of A-site troni-polyenes fit with smaller . (From Ref. 56.)...

In the previous. section we discussed the reference force field of /-PA (see Table 6.2) derived from the force field of butadiene. In the Og symmetry block, the high frequency C—H stretch is decoupled from the other modes and thus from tt electrons. We are left with three relevant Ug modes their reference and experimental frequencies are reported in Table 6.3 and, as discussed in Section II, fix the matrix and the x cd) curves in Fig. 6.4. The A matrix is written on the basis of the reference normal coordinates Q . It consequently depends on both the G and F matrices and varies with molecular or polymeric structure. The e-ph coupling constants g, thus vary even with isotopic substitution. To define coupling constants independent of mass, we use the symmetry coordinates to solve the GF problem for the reference state. In fact, diagonalization of GF gives both eigenvector matrix L in the S basis. The L matrix is used to transform Jin Eq. (12) back to the S basis ... 

The TT-electron force constants in Eq. (24) can readily be found for butadiene, which motivated the reference force field F in Table 6.2. Since is phenomenological and PPP parameters include some core effects, the precise experimental identification of 7r-electron contributions emphasized above does not extend to precise theoretical connections to PPP or other models. PPP results indicating additional delocalization beyond butadiene can simply be associated with AF(N) = F N) - F 4). The staggered bond order couples to the out-of-phase combination, 5 = (5j - 52)2 of single and double bonds, and 5 is the special ECC coordinate [22] of PA. The PPP force constants for 5 -- are... 

---