The determination of NSAR in tobacco or smokeless tobacco products has been reported several times in the literature, mainly using gas chromatography coupled with a thermal energy analyser. A different technique, liquid chromatography combined with electrospray ionisation tandem mass spectrometry (LC-ESI-MS/MS), is advanced and more convenient but involves difficulties regarding (i) the retention mechanism of the polar compound NSAR and (ii) unequal ESI-MS/MS responses of the syn and anti conformers of NSAR.

Thorough investigations by LC-ESI-MS/MS and NMR spectroscopy will be presented, as well as validation data of a method for highly sensitive NSAR quantification in smokeless tobacco products. LC-ESI-MS/MS measurements were carried out using a Liquid Separation Cell Technology Column in Hydrophilic Interaction Liquid Chromatography (HILIC) mode coupled with a triple quadrupole MS in negative ionisation mode. Sample preparation was performed according to Wu et al. (Anal. Methods 4 (2012) 3448): extraction with 2% aqueous formic acid, clean-up with solid supported liquid-liquid extraction and concentration by reconstitution of the evaporated eluate in mobile phase starting conditions. For quantification, a single-point external calibration and internal standard-correction using the isotope labeled standard NSAR-D3 was applied. NMR spectroscopy was used to assign the syn and anti conformers and to determine their ratios.

Two different LC gradient methods have been developed: one that separates the syn from the anti conformer and another that combines the two conformers. Using these two gradient methods for the determination of NSAR in standard solutions and smokeless tobacco samples, the following observations have been made: (i) the conformers show different ESI-MS/MS response, (ii) the conformer ratios differ in real samples and in standards of different ages and (iii) the conformer ratio is altered during sample preparation. As a consequence, for proper quantification the conformer ratio in the external standard solution needs to be adjusted to the same ratio as in the samples, which is accomplished by heating.

The LC-ESI-MS/MS method described has been applied to several tobacco products and has been fully validated for CORESTA Reference Product 2 (CRP2, moist snuff) and CORESTA Reference Product 3 (CRP3, dry snuff): NSAR concentrations of 31 ng/g and 63 ng/g in CRP2 and CRP3, respectively, have been determined with expanded measurement uncertainties of 21% and 16%, respectively. Recoveries of the overall method are low (average 17% for CRP2 and 9% for CRP3). Nevertheless, internal standard recoveries are consistent with analyte recoveries and therefore high intra-day and inter-day repeatabilities of the final results (5-8% RSD) are accomplished by internal standard correction. Limits of quantification are matrix dependent: 14 ng/g for CRP2 and 28 ng/g for CRP3.

The presented method is advantageous to conventional techniques in terms of selectivity, sensitivity, trueness and straightforwardness of handling. It overcomes the difficulties of LC retardation and conformer dependent ESI-MS/MS response by applying a special stationary phase and by adjusting the conformer ratio in sample and external calibration standard.