Background: Analysis of exhaled breath, especially of volatile organic compounds (VOCs

Ion mobility spectrometry is increasingly in demand for medical applications and its potential for implementation in food quality and safety or process control suggest rising use of instruments in this field as well. All those samples are commonly extremely complex and mostly humid mixtures. Therefore, pre-separation techniques have to be applied. As ion mobility spectrometers with gas-chromatographic pre-separation acquire a huge amount of data, effective data processing and automated evaluation by comparison of detected peak pattern with data bases have to be utilised. This requires accurate on-line calibration of the instruments to guarantee reproducible results, in particular with respect to identification of an analyte by determination of its ion mobility and retention time. To reduce environmental and instrumental influence, the reduced ion mobility is used. It is derived from the drift time normalised to electric field, length of the drift region and to temperature and pressure of the drift gas (traditional method). All data required for this normalisation are afflicted with a particular error and thus leading to a deviation of the calculated ion mobility value. Furthermore, this traditional method enables a calculation of the reduced ion mobility only after the measurement. To avoid those errors and to enable on-line calibration of ion mobility, an instrument specific factor is implemented generally representing all relevant variables. This factor can be determined from an initial measurement of few spectra and can thereafter be applied on the following measurement. The application of this approach obtained reproducible reduced ion mobility values for positive and negative ions over a broad drift time range and for common variation of ambient conditions as well for varying instrument conditions such as electric fields respectively drift times and in different drift gases. Moreover, the reduced ion mobility is available already during the measurements with a significantly higher reliability and accuracy which was increased to a factor of 5 compared to the traditional ion mobility determination and enables an on-line identification of analytes for the first time.

This study has developed an efficient preprocessing strategy for ion mobility spectrometry (IMS) data allowing for improved peak clarity and comparability of different measurements. Using the discrete wavelet transform for data compression and denoising, and fitting a lognormal function to the strong tailing of the reactant ion peak (RIP), enables a data reduction to 25% or less, a significant increase of the signal-to-noise ratio, and the successful elimination of the RIP tailing. The preprocessing of breath measurements obtained by coupling an IMS to a gaschromatographic column, has resulted in the desired outcome of smooth peaks lying on a common base level. These results are transferable to other applications of one-and two-dimensional separations with IMS or instrumentations generating a similar data structure.