Predicting Chemical Shifts

FIGURE 6.20 Prediction of H-NMR chemical shifts for a molecule CPG neural networks. 

FIGURE 6.21 Plot of observed chemical shifts against prediction of the neural networks for the protons of the prediction set. 

---

The program was made somewhat less convenient to use by the fact that it does not have a molecule builder. In order to predict chemical shifts, the molecular structure must be built with some other software package and then im-... 

High-level MO calculations (6-31G /MP2) give a nonclassical stmcture that is 13.6kcal more stable than foe classical stmcture and predict chemical shifts and... 

Determine the effect of basis set on the predicted chemical shifts for benzene. Compute the NMR properties for both compounds at the B3LYP/6-31G(d) geometries we computed previously. Use the HF method for your NMR calculations, with whatever form(s) of the 6-31G basis set you deem appropriate. Compare your results to those of the HF/6-311+G(2d,p) job we ran in the earlier exercise. How does the basis set effect the accuracy of the computed chemical shift for benzene ... 

As we can see, the predicted chemical shifts and coupling constants agree well with the actual values. 

The information and examples presented in this chapter should enable the reader to predict chemical shift and coupling constant values for a single fluorine substituent in virtually any possible environment in which it might be encountered. 

Structure 6.8 demonstrates a most extreme example of anisotropy. In this unusual metacyclophane, the predicted chemical shift (Table 5.8) of the methine proton that is suspended above the aromatic ring would be 1.9 ppm. In fact, the observed shift is -4 ppm, i.e., 4 ppm above TMS The discrepancy between these values is all down to the anisotropic effect of the benzene ring and the fact that the proton in question is held very close to the delocalised p electrons of the pi cloud. 

So to a large extent, 1-D 13C NMR interpretation is a case of matching observed singlets to predicted chemical shifts. These predictions can be made by reference to one of the commercially available databases that we ve mentioned, or it can be done the hard way - by a combination of looking up reference spectra of relevant analogues and using tables to predict the shifts of specific parts of your molecule (e.g., aromatic carbons). We have included some useful 13C shift data at the end of the chapter but it is by necessity, very limited. 

Advanced Chemistry Development Inc. has built a sizeable proton chemical shift database derived from published spectra (most commonly in CDCI3 solution). Their H NMR predictor programme accesses this database and allows the prediction of chemical shifts. Whilst this software takes account of geometry in calculating scalar couplings, in predicting chemical shifts it essentially treats the structure as planar. It would therefore seem doomed to failure. However, if closely related compounds, run at infinite dilution and in the same solvent, are present in the database, the conformation is implied and the results can be quite accurate. Of course, the results will not be reliable if sub-structures are not well represented within the database and the wide dispersion of errors (dependent on whether a compound is represented or not) can cause serious problems in structure confirmation (later). ACD are currently revising their strict adherence to HOSE codes for sub-structure identification and this will hopefully remove infrequent odd sub-structure selections made currently. 

Observed and Predicted Chemical Shifts for Hydrochlorinated 1.4-Pol.ydiniethyl butadiene... 

The free energies for forming products 1 and 2 were 99.68 and 99.69 kcal/mol, which are indistinguishable, considering that the uncertainty in these predictions is 4-5 kcal/mol (12). The predicted Al-NMR chemical shifts downfield from A1(N03)3 (aq) were 69.8, 69.7 ppm for product 1 and 62.7, 60.0 ppm for product 2. While the absolute values of the predicted chemical shifts are 30 ppm smaller than the observed chemical shifts, these calculations permit the assignment shown in Table 1. 

This section begins with a very brief summary of some of the technical issues associated with NMR special calculations. Subsequent subsections address the various utilities of modem methods for predicting chemical shifts and nuclear coupling constants. 

In a manner analogous to that used previously to predict chemical shifts in aromatic systems (Box 4.6), we can use substituent constants to estimate the chemical shifts of a range of compounds, e.g. substituted aromatic compounds (Box 4.16). 

Hydrogen bonding can also have a strong influence on the electronic environment of some protons. Because of this, it is sometimes difficult to predict chemical shifts. 

The values of A, given in Table 4.2 were determined for linear alkanes (standard deviation of predicted chemical shifts +0.10 ppm, B = — 2.3). 

Additional parameters Sw are required in order to calculate the chemical shift values of branched chain alkanes (standard deviation of predicted chemical shifts + 0,3 ppm, B = — 2.3 ppm) (Table 4.3). The values symbolized by Greek letters indicate the change in chemical shift due to substitution of hydrogen by a methyl group at the a to s carbon atoms. The remaining 8 correction parameters Sk, account for the effect of branching. 

Also shown in Figure 2 are the GIAO-MP2/tzp/dz values of the isotropic 13C chemical shifts. The predicted chemical shift for C2 and C3 is 255.3 ppm, which compares to the experimental value of 250 ppm. For C2, theory predicts a chemical shift of 152.8 ppm, whereas the experimental value is 148 ppm. The CH2 carbons are predicted to have isotropic chemical shifts of 50.0 ppm which compare to the measured values of 33 ppm. Finally, the methyl carbons have theoretical values of 28.9 ppm, whereas the experimental chemical shifts are 24 ppm. In all cases the theoretical values are downfield of the experimental chemical shifts. The differences are generally =5 ppm, with the exception of C3. The source of this larger difference is not clear. Still, the agreement is sufficient to verify the presence of the 1,3-dimethylcyclopentyl carbenium ion within the zeolite. 

---