Characterisation of secondary electron multiplier nonlinearity using MC-ICPMS

We have investigated the signal response characteristics of a commonly used (ETP) secondary electron multiplier (SEM) using a ThermoFinnigan Neptune multi-collector inductively coupled plasma mass spectrometer (MC-ICPMS) over a range of applied beam intensities from 10 to 5 x105 counts per second (cps). Sample switching while maintaining the same tuning parameters allows a static SEM-Faraday cup assessment of nonlinearity by MC-ICPMS, rather than a peak-switching approach as recently reported using thermal ionisation mass spectrometry (TIMS) [S. Richter, S.A. Goldberg, P.B. Mason, A.J. Traina, J.B. Schwieters, Int. J. Mass Spectrom. 206 (2001) 105]. For two SEMs of the same type (ETP) we find a count rate nonlinearity of 0.3 and 1.1% per decade of ion beam intensity variation for intensities of less than about 3 x 104 and 105 cps, respectively. Above a nominal threshold of 3 x 104 and 105 cps there is a more pronounced nonlinearity effect with an additional 0.4 and 1.6% per decade. A previous TIMS study on the same type of multiplier [S. Richter, S.A. Goldberg, P.B. Mason, A.J. Traina, J.B. Schwieters, Int. J. Mass Spectrom. 206 (2001) 105] found evidence of nonlinearity at the higher intensity range only. Although the SEM we have most rigorously tested may display an anomalously high degree of nonlinearity, we suggest that the form of behaviour is general and must be well-calibrated prior to routine high precision sample analysis. Additional tests show that after a high intensity beam was measured on the SEM of the MC-ICPMS system, the SEM yield is elevated for at least ~20 s, which can be envisaged as a memory effect related to the intensity of previously measured signals. Therefore, it is impossible to see the nonlinearity effect at low count rates using a peak jumping routine on the ICPMS because of intervening high intensity beams (e.g., 235U and 238U) applied to the SEM. This "memory" effect has important implications for MC-ICPMS measurement protocols that use multi-static or peak jump routines.