A New Algorithm for Quantitative Conversion of Non-sinusoidal Image Charge or Current Signal from Planar Electrostatic Ion Trap

Abstract

Mass resolution of Foureir transform mass spectrometer (FTMS) highly depends on the acquisition time of image charge/current signal. In order to achieve higher resolution with a shorter transient time, apart from increasing the field strength of the analyzer, the high order harmonics in the image charge signal may also be exploited. The image charge or current signal obtained from the planar electrostatic ion trap has a non-sinusoidal waveform which contains many high-order harmonic components. However, presence of high harmonics increases the complexity of spectral analysis, such as identification of a peak for its correct harmonic order, and avoiding quantitation error due to the peak overlapping from different harmonic groups. A new quantitative algorithm consists of a scoring-based peak classification and the least square fitting (SC-LSF), which has been developed to convert image charge or current signal to mass spectrum. The scoring process will go through all the peaks identified above the noise background, for assumptions that the peak belongs to a certain harmonic order. The score will go up when a relevant lower harmonic peak is confirmed. The harmonic order which achieves the highest score, is assigned to the peak so its fundamental frequency can be determined. Through the SC tests, the candidates of all fundamental frequencies are found for all possible m/z of ions. The basis signals for all possible m/z are constructed using the identified fundamental frequencies and are brought to the LSF to determine the intensities of each species. The SC-LSF algorithm was tested using simulated signal from a mixture of 48 different m/z ions. Different levels of artificial noise were added to the signal to challenge the algorithm. The results show a wide range of mass and ion numbers in the sample mixture can be accurately returned through the SC-LSF algorithm even if the transient signal is under high noise condition. LSF in frequency domain is more efficient than in time domain because a subset of frequency points may be selected, so the amount of calculation is massively reduced. In addition, the test proves that good quantitation can be achieved only when LSF is carried out using data in complex number, while fitting with magnitude data results in large errors for ions in the closed mass group. This is because those closed mass peaks are overlapping at certain low-order harmonics and the amplitudes of a FFT spectrum are not addable, thus leading to the fitting errors.

Keywords

• image charge
• fourier transform mass spectrometer (ftms)
• quantitative algorithm
• scoring-based peak classification (sc)
• least square fitting (lsf)