Recognizing and overcoming analytical error in the use of ICP-MS for the determination of cadmium in breakfast cereal and dietary supplements.

The potential effect of spectral interference on the accurate measurement of the cadmium (Cd) mass fraction in fortified breakfast cereal and a variety of dietary supplement materials using inductively coupled plasma quadrupole mass spectrometry was studied. The materials were two new standard reference materials (SRMs)--SRM 3233 Fortified Breakfast Cereal and SRM 3532 Calcium Dietary Supplement--as well as several existing materials--SRM 3258 Bitter Orange Fruit, SRM 3259 Bitter Orange Extract, SRM 3260 Bitter Orange-containing Solid Oral Dosage Form, and SRM 3280 Multivitamin/Multielement Tablets. Samples were prepared for analysis using the method of isotope dilution and measured using various operating and sample introduction configurations including standard mode, collision cell with kinetic energy discrimination mode, and standard mode with sample introduction via a desolvating nebulizer system. Three isotope pairs, (112)Cd/(111)Cd, (113)Cd/(111)Cd, and (114)Cd/(111)Cd, were measured. Cadmium mass fraction results for the unseparated samples of each material, measured using the three instrument configurations and isotope pairs, were compared to the results obtained after the matrix was removed via chemical separation using anion exchange chromatography. In four of the six materials studied, measurements using the standard mode with sample introduction via the desolvating nebulizer gave results for the unseparated samples quantified with the (112)Cd/(111)Cd isotope pair that showed a positive bias relative to the matrix-separated samples, which indicated a persistent inference at m/z112 with this configuration. Use of the standard mode, without the desolvating nebulizer, also gave results that showed a positive bias for the unseparated samples quantified with the (112)Cd/(111)Cd isotope pair in three of the materials studied. Collision cell/kinetic energy discrimination mode, however, was very effective for reducing spectral interference for Cd in all of the materials and isotope pairs studied, except in the multivitamin/multielement matrix (SRM 3280) where the large corrections for known isobaric interferences or unidentified interferences compromised the accuracy. For SRM 3280, matrix separation provided the best method to achieve accurate measurement of Cd.