A prototype inductively coupled plasma time-of-flight mass spectrometer (ICPTOFMS) for time resolved measurements of transient signals in the microsecond regime is described in this work. Analytical figures of merit for the prototype are given for both liquid nebulization and single droplet introduction and are compared to a conventional quadrupole-based ICPMS using the same ICP source and vacuum interface. Quasi-simultaneous detection at a time resolution of 33 μs of the prototype ICPTOFMS allows multi-isotope monitoring of short signals (200–500 μs duration) generated from individual droplets and particles. The capabilities of the instrument for the analysis of single nanoparticles are studied using microdroplets consisting of a multi-element standard solution and containing 114 nm Au particles. The detection efficiencies for Ag and Au, calculated from the response of individual droplets and particles, are similar to those of the quadrupole-based instrument and amount to 1.3 × 10⁻⁶ ions per atom and 3.1 × 10⁻⁶ ions per atom, respectively. The sizes of the smallest detectable Ag, Au and U metallic nanoparticles are estimated to be 46 nm, 32 nm and 22 nm, respectively. Furthermore, time shifts of the signals of different elements within single droplets were observed. These new results demonstrate the advantage of the temporal resolution of the instrument for studying processes taking place in the plasma on the μs-time scale.

Signal to noise ratios of transient signals depend on the signal duration and are usually better for shorter signals. Poisson distribution is the dominating noise source for mass spectrometers at low count numbers. The IUPAC concept of ‘detection limit’ is used as a measure of the noise level and three different ways of detection limit calculations are compared and discussed. Assumption of Normal distribution leads to underestimation of the detection limit compared to exact Poisson based calculations. A simple formula given by IUPAC bases detection limit on the standard deviation of the blank but introduces a term to correct for the difference between Normal and Poisson distributions. Results of this formula are comparable to exact Poisson calculations. Signal to noise in transient signals was evaluated. It was found that an optimum duration for transient signals exists and that it depends on the level of the instrument background.