Analytical procedures have been developed for the reliable determination of 19 trace elements (Ag, Al, Ba, Bi, Cd, Co, Cr, Cu, Fe, Mn, Pb, Rb, Sb, Sc, Sr, Tl, U, V, Zn) in ice samples at pgg −1 and fgg −1 concentrations using ICP-sector field mass spectrometry (ICP-SMS). Concentrations of most elements in the high purity water and doubly distilled HNO 3 employed were distinctly lower than previously reported values. The accuracy of the results was carefully evaluated using the certified water reference material SLRS-4. Contributions of unwanted trace elements due to acidification of the ice samples (0.5% HNO 3 ) to the total element budget amounted to only 0.001pgg −1 for Bi, 0.34pgg −1 for Cr, 0.2pgg −1 for Fe, 0.004pgg −1 for Pb, 0.00015pgg −1 for U and 0.0025pgg −1 for V: compared to the concentrations of the metals in ice these are negligible. The use of a detergent (0.05%) in the rinsing solution (0.5% HNO 3 ), helped to reduce memory effects by 59–98%, depending on the element considered; this resulted in shorter washing times between samples (i.e. 1min) and improved analysis time. Adopting strict clean room procedures, the detection limit for Pb (0.06pgg −1 ) is a factor of ten lower than the current state-of-the-art. Compared to previous studies, the improved LODs obtained here for other trace elements amount to 2× (Ag), 4× (Sb), 5× (Ba), 6× (Cu, Mn, U), 9× (Bi), 13× (Cd), 18× (Fe) and 21× (V). The developed analytical protocols were successfully applied to the determination of selected trace elements in age-dated ice samples from the Canadian High Arctic. The toxic trace element Tl (median: 0.16pgg −1 ; range: 0.03–1.32pgg −1 ) and the lithogenic reference element Sc (0.53pgg −1 ; 0.06–2.9pgg −1 ) have been determined in a polar ice core for the first time.