Send to

Choose Destination
Anal Bioanal Chem. 2016 Jul;408(19):5075-87. doi: 10.1007/s00216-016-9509-9. Epub 2016 Apr 19.

Drift correction of the dissolved signal in single particle ICPMS.

Author information

Department of Soil and Environment, Swedish University of Agricultural Sciences, Box 7014, 75007, Uppsala, Sweden.
Department of Chemistry and Molecular Biology, University of Gothenburg, Kemivägen 10, 41296, Göteborg, Sweden.
Water Environment Technology, Department of Civil and Environmental Engineering, Chalmers University of Technology, 412 96, Göteborg, Sweden.


A method is presented where drift, the random fluctuation of the signal intensity, is compensated for based on the estimation of the drift function by a moving average. It was shown using single particle ICPMS (spICPMS) measurements of 10 and 60 nm Au NPs that drift reduces accuracy of spICPMS analysis at the calibration stage and during calculations of the particle size distribution (PSD), but that the present method can again correct the average signal intensity as well as the signal distribution of particle-containing samples skewed by drift. Moreover, deconvolution, a method that models signal distributions of dissolved signals, fails in some cases when using standards and samples affected by drift, but the present method was shown to improve accuracy again. Relatively high particle signals have to be removed prior to drift correction in this procedure, which was done using a 3 × sigma method, and the signals are treated separately and added again. The method can also correct for flicker noise that increases when signal intensity is increased because of drift. The accuracy was improved in many cases when flicker correction was used, but when accurate results were obtained despite drift, the correction procedures did not reduce accuracy. The procedure may be useful to extract results from experimental runs that would otherwise have to be run again. Graphical Abstract A method is presented where a spICP-MS signal affected by drift (left) is corrected (right) by adjusting the local (moving) averages (green) and standard deviations (purple) to the respective values at a reference time (red). In combination with removing particle events (blue) in the case of calibration standards, this method is shown to obtain particle size distributions where that would otherwise be impossible, even when the deconvolution method is used to discriminate dissolved and particle signals.


Chemometrics/statistics; Mass spectrometry/ICPMS; Metals/heavy metals; Nanoparticles/nanotechnology


Supplemental Content

Full text links

Icon for Springer
Loading ...
Support Center