analytica chimica acta 611 ( 2 0 0 8 ) 134–142 available at www.sciencedirect.com journal homepage: www.elsevier.com/locate/aca Optimisation and critical evaluation of a collision cell technology ICP-MS system for the determination of arsenic in foodstuffs of animal origin Vincent Dufailly, Laurent No¨ el, Thierry Gu´ erin * Unit´ e des Contaminants Inorganiques et Min´ eraux de l’Environnement, AFSSA-LERQAP 23, Avenue du G. de Gaulle, F-94706 Maisons-Alfort Cedex, France article info Article history: Received 11 September 2007 Received in revised form 8 February 2008 Accepted 9 February 2008 Published on line 16 February 2008 Keywords: Inductively coupled plasma mass spectrometry Collision cell Foodstuffs Arsenic Interferences abstract The determination of arsenic ( 75 As) was studied using an ICP-MS equipped with collision cell technology (CCT). Different mixtures of gases (He and H 2 ) were tested using HCl conditions and a He flow rate of 4 mL min -1 was found to be suitable for the removal of the poly-atomic spectral interference [ 40 Ar 35 Cl] + . Trueness of the optimised method has been evaluated in both standard and CCT modes on six certified reference materials in foodstuffs of animal origin and on three external proficiency testing schemes (FAPAS). The results obtained gener- ally coincided with the certified values, except for CCT mode in some categories of samples (meat, mussels and milk powder), for which a positive bias on results was observed due to the formation of poly-atomic interferences within the collision cell. The main interfer- ences were studied and their contributions estimated. [ 58 Fe 16 O 1 H] + and [ 74 Ge 1 H] + were the most significant interferences formed in the cell. Finally, different parameters (e.g. hexapole and quadrupole bias voltage, nebuliser gas flow) were optimised to try to attenuate these interferences. © 2008 Published by Elsevier B.V. 1. Introduction Inductively coupled plasma mass spectrometry (ICP-MS) has rapidly become an established method of trace multi-element analysis in various matrices such as biological fluids, food and beverages [1–10]. Yet, like any sensitive analytical techniques, it still has some limitations such as spectral interferences. Most of these interferences are predictable and can be min- imised in many cases by the use of alternative isotopes with lower natural abundances and/or by mathematical correction equations. However, not all determinations are straightfor- ward, in particular the determination of elements which undergo major poly-atomic interferences generated by the ∗ Corresponding author. Tel.: +33 1 49 77 27 11; fax: +33 1 49 77 26 50. E-mail address: t.guerin@afssa.fr (T. Gu´ erin). plasma gas (Ar), matrix components (e.g. O, C, N, Cl) or solvent- acid (HNO 3 , HCl) [1]. One of the main poly-atomic ions, [ 40 Ar 35 Cl] + , has nearly the same mass-to-charge ratio (m/z 75) as monoisotopic arsenic (As), and could interfere with its determination in samples with significant amounts of chlorine. The most com- monly used technique to reduce the [ 40 Ar 35 Cl] + interference in quadrupole ICP-MS is the well-known mathematical correc- tion equation [2,11–13]. However, this correction formula may introduce errors in the case of a routine method of analysis if sample matrix components vary, and can be particularly unreliable for low concentrations of arsenic [11]. Thus, some recent applications have been optimised on quadrupole ICP- 0003-2670/$ – see front matter © 2008 Published by Elsevier B.V. doi:10.1016/j.aca.2008.02.018