Anal. Chem. 2006, 78, 4474-4481
Hadamard Transform Ion Mobility Spectrometry Andrew W. Szumlas, Steven J. Ray, and Gary M. Hieftje*
Indiana University, Department of Chemistry, Bloomington, Indiana 47405
A detection scheme that makes use of the Hadamard transform has been employed with an atmosphericpressure ion mobility spectrometer fitted with an electrospray ionization source. The Hadamard transform was implemented through the use of a linear-feedback shift register to produce a pseudorandom sequence of 1023 points. This pseudorandom sequence was applied to the ion gate of the spectrometer, and deconvolution of the ion signal was accomplished by the Hadamard transform to reconstruct the mobility spectrum. Ion mobility spectra were collected in both a conventional and Hadamard mode, with comparisons made between the two approaches. Initial results exhibited low spectral definition, so an oversampling technique was applied to increase the number of data points across each analyte spectral peak. The use of the Hadamard transform increases the duty cycle of the instrument to 50% and results in a roughly 5-fold enhancement of the signal-to-noise ratio with a negligible loss of instrument resolution. It is also shown that any potential multiplex disadvantage, which limits the attractiveness of some high-throughput techniques, is not a limiting factor in this new implementation. Ion mobility spectrometry (IMS) has recently received increased attention largely because of the need for fast, selective, and sensitive detectors in homeland security and defense applications.1 Growth in the development of ion mobility instrumentation has been driven mainly by two needs: devices in airports to screen for narcotics and explosives and portable instruments used by the military for the detection of compounds such as chemical warfare agents.2 The widespread adoption of IMS instruments is partially because of their ability to be made compact and highly portable, while at the same time, they produce excellent detection levels and have a reasonable degree of selectivity. For these reasons, IMS is often employed as an initial screening method, where qualitative detection of harmful or illicit compounds is the focus. The simple physical requirements of IMS instruments, which include a lack of vacuum pumps, small drift-tube sizes, and uncomplicated electronics, are a highlight of the technique’s utility. However, the same physical simplicity can also limit the merit for these instruments. For example, since the IMS drift cell operates under a nitrogen or similar gas atmosphere and often at * To whom correspondence should be addressed. Fax: 812-855-0958. Email:
[email protected]. (1) Smith, W. D. Anal. Chem. 2002, 5, 462A-466A. (2) Eiceman, G. A.; Stone, J. A. Anal. Chem. 2004, 76, 390A-397A.
4474 Analytical Chemistry, Vol. 78, No. 13, July 1, 2006
atmospheric pressure, simple Faraday (charge-collecting) detectors are commonly used. With the help of high amplification, these collectors provide accurate measurement of ion signals, yet lag behind other ion detectors, such as those used in mass spectrometry which produce multiple electrons per analyte ion and allow individual ions to be counted. Another limitation of ion mobility spectrometry is modest selectivity because the analyte ions are distinguished strictly by differences in their collisional crosssections in the gas phase. The resolution of some instruments has reached 100-150 through the use of specialized drift tubes,3 but since the length of portable drift-tube devices should be kept small, their resolution suffers accordingly. The sensitivity exhibited by IMS, limited in part by detector technology, has led other researchers to examine novel methods that might provide lower detection limits. One such method, introduced by Knorr et al.,4 employed a Fourier transform (FT) to increase the instrument duty cycle. Duty cycle (DC) is a measure of the fraction of ions that reach the detector out of the whole set of ions presented to the instrument. Improvement in DC leads directly to signal-to-noise (S/N) improvements (I) as
IS/N )
x
DCimproved DCprevious
(1)
Since the FT-IMS approach predicted a DC of 25% in comparison to the
