The quite popular, simple but imperfect method of referencing NMR spectra to residual 1H and 13C signals of TMS-free deuterated organic solvents (referred to as Method A) is critically discussed for six commonly used NMR solvents with respect to their δH and δC data that exist in the literature. Taking into account the most reliable data, it was possible to recommend ‘best’ δX values for such secondary internal standards. The position of these reference points on the δ scale strongly depends on the concentration and type of analyte under study and the solvent medium used. For some solvents, chemically induced shifts (CISs) of residual 1H lines were considered, also taking into account the formation of 1:1 molecular complexes (for CDCl3). Typical potential errors that can occur as a result of improper application of Method A are considered in detail. An overview of all found δX values adopted by users of this method revealed a discrepancy of up to 1.9 ppm in δC reported for CDCl3, most likely caused by the CIS mentioned above. The drawbacks of Method A are discussed in relation to the classical use of an internal standard (Method B), two ‘instrumental’ schemes in which Method A is often implicitly applied, that is, the default Method C using 2H lock frequencies and Method D based on Ξ values, recommended by the IUPAC but only occasionally used for 1H/13C spectra, and external referencing (Method E). Analysis of current needs and opportunities for NMR spectrometers led to the conclusion that, for the most accurate application of Method A, it is necessary to (a) use dilute solutions in a single NMR solvent and (b) to report δX data applied for the reference 1H/13C signals to the nearest 0.001/0.01 ppm to ensure the precise characterization of new synthesized or isolated organic systems, especially those with complex or unexpected structures. However, the use of TMS in Method B is strongly recommended in all such cases.
