International Journal for Ion Mobility Spectrometry

, Volume 12, Issue 2, pp 59–63

Empirical prediction of reduced ion mobilities of secondary alcohols


    • ISAS — Institute for Analytical Sciences
  • Jörg Ingo Baumbach
    • ISAS — Institute for Analytical Sciences
  • Wolfgang Vautz
    • ISAS — Institute for Analytical Sciences
Original Research

DOI: 10.1007/s12127-009-0017-x

Cite this article as:
Hariharan, C., Ingo Baumbach, J. & Vautz, W. Int. J. Ion Mobil. Spec. (2009) 12: 59. doi:10.1007/s12127-009-0017-x


Ion Mobility Spectrometry is a powerful method for the rapid identification of gas-phase analytes and finds its usage in various fields including the sensitive analysis of extremely complex and humid mixtures such as human breath when additional pre-separation techniques are applied. The output data from an ion mobility spectrometer (IMS), equipped with a Multi-Capillary Column (MCC) for pre-separation, is a chromatogram of the signal intensity versus a particular retention time and a specific reduced ion mobility which are the characteristics of the detected analyte. Hence, it is important to have a database of analytes with both the values for comparison and identification of peaks in any IMS chromatogram. Commonly, such databases are collected by measurements of reference analytes. It is obvious that a prognosis of the values, without the time consuming and costly reference measurements, would be a considerable facilitation for a preliminary identification of unknowns and development of databases. In this study, a correlation between the reduced ion mobilities and the number of carbon atoms was found for secondary alcohols. The correlation was then used to predict the reduced ion mobilities of other analytes in the same homologous series. To verify the accuracy of the prognosis, the analytes were measured individually using a 63Ni-MCC-IMS and compared to the predicted values. The results of the prognosis show an accuracy higher than 99.5%.


Ion mobility spectrometryDatabaseSecondary alcoholsEmpirical predictionReduced ion mobilityHomologous series


Over the past few decades, ion-mobility spectrometry has evolved into an inexpensive and powerful technique for the detection of trace compounds in the lower ng/L (ppbv) down to pg/L (pptv) range, for direct monitoring of specific compound classes such as chemical warfare agents and drugs of abuse [1, 2]. However, in recent years, ion mobility spectrometers are increasingly in demand for new applications specifically on biological samples (cells, fungi, bacteria) [312], in medicine (diagnosis, therapy and medication control e.g. from breath analysis) [3, 4, 13, 14] and process control [1522]. For such applications, IMS measurements faces challenges such as humid and rather complex samples, requirement of a specific sampling procedure adapted to the application, fast pre-separation techniques like multi-capillary columns and most importantly, suitable data processing techniques which include databases of relevant analytes for automatic characterisation of the signals detected in an IMS chromatogram [2328], and different data pre-processing steps [2932].

Till recently, IMS have been used to detect specific target analytes with known reduced ion mobility. For the analysis of complex mixtures, the reduced mobility alone will not be sufficient for the identification of analytes in the mixture. Several analytes have similar or even the same mobility [4]. Hence, additional rapid pre-separation techniques are applied. The time taken by an analyte to elute out of the MCC is termed as the retention time (tR) of the analyte. Thus, for every analyte, there is a specific reduced ion mobility value and a retention time which are the characteristics of a particular analyte at a specific temperature, column length and flow. Providing this database of relevant analytes for every analysis enables the identification of compounds in an unknown complex mixture.

During the analysis of complex mixtures, e.g. human breath, volatile organic compounds from bacteria etc., there are many unknown peaks in the IMS chromatogram. Identifying them is one of the major problems. There are a huge number of analytes available from which one could select, measure and identify the unknown peak. Most applicable solution is the additional sampling on adsorption materials and analysis by GC/MS. From such analysis, substances possibly responsible for the unknown signals could be proposed. This proposal has to be validated by IMS measurements of the reference analyte. Obviously, such a procedure is time consuming and expensive. Thus, additional tools for the identification of unknown analytes would be very helpful. There have been earlier attempts to calculate the ion mobility of analytes directly from the molecular structure using computational neural networks and multiple linear regression analysis [33, 34]. But, these methods are seemingly complex when compared to the prediction method proposed in this study and hence cannot be used for online characterisation of complex mixtures.

In this study, a rather simple method to see trends with the secondary alcohols between the inverse reduced ion mobility and the number of carbon atoms in the analyte is used to predict the reduced ion mobilities of other unmeasured analytes in that particular homologous series directly.


Ion mobility spectrometry

In the drift tube—under the influence of an external electric field and with a particular drift gas—ions of different masses and /or structure reach different velocities and thus, get separated [2]. The quotient of ion velocity and electric field strength is referred to as ion mobility [35]. The mobility, K, is calculated using the measured drift time, tD, of an ion through a specified drift length, lD, under a known electric field, E, using the equation below [1, 2].
$$ v_d = \frac{{l_D }}{{t_D }} $$
$$ v_D = K*E $$
This ion mobility value is normalised to standard gas density, 2.687 × 1019 molecules/cm, corresponding to 273 degree Kelvin and 101325 Pascal, and reported as the reduced ion mobility, K0 [3537].
$$ K_0 = K\left( {\frac{P}{101325}} \right)\left( {\frac{273}{K}} \right) $$
At the molecular level, K is dependent on several factors and can be described by the equation,
$$ K = \left( {\frac{3q}{16N}} \right)\left( {\sqrt {\frac{{2\pi }}{kT}} } \right)\left( {\sqrt {\frac{m + M}{mM}} } \right)\left( {\frac{1}{\Omega }} \right) $$

In equation 4, q is the ionic charge (1.602 × 10-19 C), N; the number density of the drift gas, k; the Boltzmann constant (1.381 × 10-23 JK-1), T; the temperature in Kelvin, m; the ion mass, M; the mass of the drift gas and Ω; the ion collision cross section [1, 34, 35, 37, 38].

The basic working principles of an IMS and the details of the IMS used in this study have been described in detail in previously published articles [1, 2, 5, 39, 40]. A multi-capillary column (MCC OV-5; Multichrom, Novosibirsk, Russia) was used for rapid pre-separation and was operated at a constant temperature of 40°C. The experimental parameters of the IMS used in this study are summarised in Table 1.
Table 1

Experimental parameters of 63Ni-MCC-IMS

Experimental Parameters

Ionisation source

β radiation (63Ni, 550 MBq)

Grid opening time

300 µs

Spectral length / Sample interval

100 ms

Spectral resolution

40 kHz

Drift length

120 mm

Electric field intensity

310 V/cm

Drift gas & carrier gas

Synthetic air

Drift gas flow

100 mL/min

Carrier gas flow

150 mL/min


MCC OV-5, 40°C (constant), 20 cm

As the drift time tD—which is proportional to the inverse reduced ion mobility—is the measured parameter, the inverse ion reduced mobility values (in are used in the following study for the visualisation of data and for correlation of predicted and measured ion mobilities.

All reference analytes were obtained from Sigma Aldrich (puriss. p.a.). The calibrated gases were provided by a calibration gas generator (HovaCAL 3834SO-VOC, IAS, Frankfurt, Germany). 2-Nonanone was always included in the calibration gases and was used as an internal standard for the retention time together with the reactant ion peak (RIP) for the ion mobility.

Results and discussion

From the database of analytes available at ISAS—Institute of Analytical Sciences, Dortmund, Germany, the homologous series of secondary alcohols was chosen for this study. The ion mobilities of 2-Hexanol, 2-Heptanol, 2-Octanol and 2-Undecanol were previously measured and the values are given in Table 2. From these data, the inverse reduced mobility values were plotted against the number of carbon atoms (Fig. 1). The data points were fit linearly and the corresponding equation is indicated.
Table 2

Measured, predicted and validated inverse reduced ion mobility values of secondary alcohols



No. of Carbon atoms

Measured 1/K0in

Measured K0in cm².V-1s-1

std. dev. in

Predicted 1/K0in

Accuracy in %


























Predicted & validated




























Fig. 1

Correlation of the inverse reduced ion mobility and the number of carbon atoms of the four secondary alcohols. The linear fit of the points yields the function y = 0.0334x + 0.4303 with a correlation coefficient of 0.9999

Using the linear equation obtained from Fig. 1, y = 0.0334x+ 0.4303, the inverse reduced ion mobility values of other analytes in the same homologous series were empirically predicted as given in Table 2. To validate the quality of this empirical prediction, the analytes prognosed were individually measured with a 63Ni-MCC-IMS using the same experimental parameters given in Table 1. The reference gases were provided by the calibration gas generator stated previously. The inverse ion mobilities obtained from the spectra (Fig. 2) are given in Table 2. The measurements were carried out several times with a standard deviation of less than 0.0005 The accuracies of the empirical prediction of inverse reduced ion mobilities of the analytes, when compared to the measured ones, were greater than 99.5% (Fig. 3).
Fig. 2

Inverse reduced ion mobility values of the 4 predicted secondary alcohols together with the internal standard (2-Nonanone). The spectra were obtained at the retention time of the corresponding monomer peak maxima
Fig. 3

Predicted and measured inverse reduced ion mobility values of measured and predicted secondary alcohols

Conclusion and outlook

A linear correlation between the number of carbon atoms with the inverse reduced ion mobilities of secondary alcohols was used to predict inverse ion mobility values of other analytes in the same homologous series. This empirical prediction method proves that the inverse reduced mobilities of analytes in the IMS can be predicted very accurately from known values of the analytes in the same homologous series. More work has to be done to provide a clear insight into the reasons behind the trend within a homologous series. Furthermore, this empirical prediction will be extended to other homologous series.

Presently, this prognosis is being extended to other homologous series in order to develop a generalised equation for the prediction of the reduced ion mobility with different variables for every series depending on the molecular structure of the analyte. This would also lead to an extension of analytes’ database with the predicted values from which the responsible analyte for any unknown peak in an IMS chromatogram could be identified. However, the theoretical correlation of the presented approach to the molecular structure of the analyte and to the behaviour of its ions in the drift gas under the influence of the electric field requires further investigations.

Such predictions will assist peak characterisations using Breit-Wigner-Functions and simulations of the shape of peaks in 3-D plots of IMS-chromatograms and allow forecast of peak positions within the drift time-retention time-intensity space. Furthermore, it could be helpful to reduce false alarms and to separate overlapping signals [article submitted for publication].


The financial support of the Bundesministerium für Bildung und Forschung and the Ministerium für Wissenschaft und Forschung des Landes Nordrhein-Westfalen is gratefully acknowledged. The dedicated work of Luzia Seifert and Susanne Krois, both technicians at ISAS, was indispensable for the success of the investigations. The work was funded partly by the project BAMOD (Breath-gas analysis for molecular-oriented detection of minimal diseases) of the European Union (LSHC-CT-2005-019031) and the high-tech strategy funds of the Federal Republic of Germany (Project Metabolit-01SF0716).

Copyright information

© Springer-Verlag 2009