Thoughts on the ACD/ P-31 NMR Database and Prediction Software Program

P-31 NMR is an absolutely essential tool in modern phosphorus chemistry, and it has aided countless investigations in this area. Numerous applications of the technique are in practice. With the modern high resolution instruments, peak separation of a few hundredths of a ppm can be achieved for mixtures of phosphorus compounds. Even complex mixtures such as extracts of biophosphates or biophosphonates give well-resolved spectra. Furthermore, cases of identical chemical shifts arising from different compounds at this level of resolution are extremely rare, allowing confident assignment of peaks when a match with a known compound is obtained. Since signal intensity can be related to concentration by proper calibration procedures, quantitative analysis is possible for reaction products, bioextracts, metabolic products from pharmaceuticals and agrochemicals, etc. Variable temperature measurements can be used in studying equilibrium processes, such as those of conformational, fluxional or tautomeric nature. Measurements for the applications mentioned are generally made on solutions, but with modern techniques solid state samples or solids with phosphorus species bonded on the surface can be readily studied.

But essential in any P-31 NMR study is an understanding of the phosphorus chemical shift as a function of molecular structure. While progress has been made in recent years in the calculation ab initio of P-31 shifts (recently reviewed by Chesnut and Quin¹), the time-tested approach of comparing shifts of an unknown to a reference collection of shifts remains a standard way of operation. The shifts of different classes of phosphorus functions fall into definite ranges (although with some overlap), and the association of a shift with a type of phosphorus function can frequently be made with confidence. But getting at subtle shift effects of molecular structure within a functional group family in general requires an empirical approach (as outlined in a recent book²) which depends on the availability of a reference collection of spectra. In years past, phosphorus chemists made extensive use of the tables of data obtained by the older continuous wave technique (with P-H coupling present) that were published in 1967 by Van Wazer and associates³. More recently (1991) John Tebby edited an entire handbook⁴ of P-31 shifts which for the most part reported proton-decoupled data obtained from the more precise Fourier Transform method. But even the Tebby tables are becoming dated, and it is of great significance that computer databases of P-31 NMR shifts and couplings are being developed.

Click on picture to get molecules in ChemSketch format.

One of these, by Advanced Chemical Development (ACD), contains at this writing 23,400 P-31 shifts, with literature references and the correct IUPAC names. P-H and P-C coupling data are included, a great advance over the printed tables. Many of the shifts will of course duplicate those of the the printed tables, but the database contains some shifts for compounds published after the closing of the Tebby tables. Additions are continually made to the database, and regular updates are to be provided to subscribers. The ease of searching the database, and the speed of locating particular compounds with their shifts, is remarkable. The search can be accomplished by several methods, such as drawing a structure, using a correct name, calling up a selected range of chemical shifts for all types of functions, or a range limited to a particular function by a substructure search. The database can be an aid also in showing the structure of an unidentified laboratory product. In an example from my own recent work (with A.S. Ionkin), the flash vacuum pyrolysis of the cyclic trimer (ArO-P=S)₃ was performed so as to generate the previously unknown monomeric form. None of this species could be detected and the major product had a P-31 spectrum of a totally unexpected nature, with one P-P coupled (69.3 Hz) signal in the well upfield region of d -129.3, the other downfield at d +69.3. Examining the database for chemical shifts in the d -120 to -130 region would immediately reveal the answer; the spectrum is consistent with that found for the well-known but totally unexpected phosphorus sesquisulfide (P₄S₃), a product of extensive and complicated degradation reactions. The database in this case provided 7 references to published spectra.

Click on picture to get molecules in ChemSketch format.

Another advance made possible by computer technology is the rapid computation of the approximate chemical shift to be expected for a particular structure not found in the database. The ACD program for doing this depends on shifts already in the database for closely related structures; this gives an empirical value for the shift, not to be confused with shifts from the fundamental ab initio calculations. For the most part, unless one is dealing with molecules with unusual steric or electronic properties, the empirically calculated spectra are in reasonable agreement (generally a few ppm) with the recorded value. When pronounced deviations do occur, the result becomes of value in detecting the presence of unusual influences, being especially prominent among heterocyclic phosphorus compounds. A prime example from my own work is the case of the 7-phosphanorbornene structure. With a P-methyl substituent, the prediction gives a value of d 11.97, but fails to detect a difference between syn and anti isomers. But real molecules with this structural feature give remarkably downfield shifts of about d +100 for the syn isomer and d +25 for the anti isomer. Clearly a strong unexpected effect is present in the syn isomer, and this was associated with MO characteristics of the structure¹. Thus, while the prediction can be of value in confirming that an experimental value is consistent with that of a structural type, it can also reveal the presence of an unusual influence on chemical shifts.

I consider the new computer programs that provide the database and shift prediction possibilities for P-31 NMR to be a major step forward in practical organophosphorus chemistry.

References:

1. D.B. Chesnut and L.D. Quin, in M. Hargittai and I. Hargittai, eds., Advances in Molecular Structure Research, Vol. 5, JAI Press Inc, Stamford, CT, pp. 189-222.
2. L.D. Quin, A Guide to Organophosphorus Chemistry, John Wiley and Sons, New York, 2000, Chapter 6.
3. M.M. Crutchfield, C.H. Dungan, J.H. Letcher, V. Mark and J.R. Van Wazer, Topics in Phosphorus Chemistry 5, 1967, pp. 227-457.
4. J.C. Tebby, ed., Handbook of Phosphorus-31 Nuclear Magnetic Resonance Data, CRC Press, Boca Raton, FL, 1991.