Prediction of spectra

To identify the main methods and tools available for the computer prediction of spectra from the molecular structure, and for automatic structure elucidation from spectral data... 

To realize that a proper representation of the molecular structure is crucial for the prediction of spectra... 

Perhaps more valuable over time than the quantitative predictions of spectra, structural parameters, and relative enthalpies and entropies of RIs, which can be obtained from electronic structure calculations, are the qualitative models of the electronic structures and reactivities of RIs that emerge from the computational results. Any model, to be successful, must do two things. 

Near the line centers, the spectral functions have sometimes been approximated by a Lorentzian. The far wings, on the other hand, may be approximated by exponential functions as Fig. 3.2 might suggest. However, better model profiles exist see Chapters 5 and 6 [421, 102, 320], Model profiles have been useful for fitting experimental spectra, for an extrapolation of measured profiles to lower or higher frequencies (which is often needed for the determination of spectral moments) and for a prediction of spectra at temperatures for which no measurements exist. We note that van der Waals dimer structures (which appear at low frequencies and low pressures) modify the Lorentzian-like appearance more or less, as we will see. 

The question arises whether collision-induced profiles may perhaps also be modeled by these or other simple functions, perhaps under circumstances to be defined. Such model functions would be of interest for the analysis of measured spectra, for empirical predictions of spectra at temperatures other than those of the measurements, for frequency extrapolations as may be necessary for accurate determination of spectral moments, etc. 

Fig. 10 Root mean square errors for prediction of spectra and concentration of pyrene using PLS1 as successive number of components are employed.

Calculations on transition metal complexes are reviewed. The limitations of ab initio calculations are shown to make predictions of spectra, bond lengths, bond energies, and spin state very difficult. Calculations on bisammineporphinatoiron, hydridocobaltcarbonyIs, and scandium carbonyl are reviewed. 

Examples are the Tripos force field (22), the COSMIC force field (23), and that of White and Bovill (24), which uses only two atom types, those at the end of the bond to parameterize the torsional potential rather than the four types of the atoms used to define the torsional angle. One has only to consider the number of combinations of 20 atom subtypes taken four at time (160,000) versus two at a time (400) to understand the explosion of parameters that occurs with increased atom sub-types. The simplifying assumption in parameterization of the torsional potential reduces to some extent the quality of the results (25), but allows the use of the simplified force fields (22) in many situations where other force fields would lack appropriate parameters. The situation can become complicated, however. For example, the amide bond is normally represented by one set of parameters, whether the configuration is cis or trans. Experiments data are quite compelling that the electronic state is different between the two configurations, and different parameter sets should be used for accurate results (Fig. 3.1). Only AM-BER/OPLS currently distinguishes between these two conformational states (26). Certainly, the limited parameterization of simplified force fields would not allow accurate prediction of spectra that is more reflective of the dynamic behavior of the molecule. 

In the context of analytical databases, the simulation or prediction of spectra from molecular descriptors is of interest. Of course, the descriptors can be correlated with any kind of properties or activities a molecule might have. This is then called QSPR... 

In conclusion, one can again reaffirm what already has been established by many workers in the field, namely, that the propagator theory is an appropriate and practical approach to the interpretation and prediction of spectra. The results presented here also show that in order to contain truly quantitative agreement with experiment, it is necessary to consider electron propagator theory at the third and partial fourth order and to also be able to accommodate larger basis sets. 

B. W. Wabuyele and P. de B. Harrington, Chemom. Intel Lab. Syst., 29, 51 (1995). Quantitative Comparison of Bidirectional and Optimal Associative Memories for Background Prediction of Spectra. 

ABSTRACT. Predictions are presented of spectra for excitation of the van der Waals rovibrational modes in ArHCl, ArHCN, H2DF, ArOH and NeC2H4. For ArHCN, H2DF and ArOH the potential energy surfaces used in the spectral computations have been obtained from CEPA calculations with large basis sets. Comparisons with experiment illustrate the power and usefulness of ab initio methods in predicting spectra for van der Waals molecules. The results also demonstrate that predictions of spectra can now be made for van der Waals molecules more complicated than the complexes of atoms with closed-shell diatomics. 

When this convention is followed, remarkably good predictions of spectra are obtained for even AHs. Indeed, the results are usually much better than those given by the standard Hiickel method. Table 6.1 lists structures and calculated and observed wavelengths for the first tt tt transitions for a wide range of benzenoid hydrocarbons. Values calculated by the Hiickel method are included for comparison. 

There are theoretical procedures in use in the literature which have led to accepted definitions of parameters characteristic for a molecule under consideration. The underlying expressions will be given here. It is aimed at that the reader is provided with sufficient information to calculate, at least in principle, the pertinent energetic stmctures of symmetric-top molecules with the aid of the parameter listings given in this sub-volume. The knowledge of the appropriate selection rules is then necessary for the prediction of spectra. 

The development of computational approaches to vibrational spectra of systems in the condensed phase requires the analysis of the physics of the solvated system, thus introducing in the model ways for treating the interaction between the system and its surrounding. This is a general requirement for a procedure to be successful in the description (and further prediction) of spectra it is in fact well known that all spectral features (frequencies, intensities, and bandshape) are hugely influenced by the interactions between the system and the environment. 

---