ANALYST FULL PAPER THE www.rsc.org/analyst Micro-machined planar field asymmetric ion mobility spectrometer as a gas chromatographic detector G. A. Eiceman,* a E. G. Nazarov, a R. A. Miller, b E. V. Krylov a and A. M. Zapata c a Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, NM 88003, USA b Sionex Corporation, Wellesley Hills, MA 02481, USA c Draper Laboratory, Cambridge, MA, USA Received 18th December 2001, Accepted 1st March 2002 First published as an Advance Article on the web 20th March 2002 A planar high field asymmetric waveform ion mobility spectrometer (PFAIMS) with a micro-machined drift tube was characterized as a detector for capillary gas chromatography. The performance of the PFAIMS was compared directly to that of a flame ionization detector (FID) for the separation of a ketone mixture from butanone to decanone. Effluent from the column was continuously sampled by the detector and mobility scans could be obtained throughout the chromatographic analysis providing chemical information in mobility scans orthogonal to retention time. Limits of detection were approximately 1 ng for measurement of positive ions and were comparable or slightly better than those for the FID. Direct comparison of calibration curves for the FAIMS and the FID was possible over four orders of magnitude with a semi-log plot. The concentration dependence of the PFAIMS mobility scans showed the dependence between ion intensity and ion clustering, evident in other mobility spectrometers and atmospheric pressure ionization technologies. Ions were identified using mass spectrometry as the protonated monomer and the proton bound dimer of the ketones. Residence time for column effluent in the PFAIMS was calculated as ~ 1 ms and a 36% increase in extra-column broadening versus the FID occurred with the PFAIMS. Introduction Detectors in gas chromatographic separations have long supplemented the chromatographic performance of columns by providing additional information related to a sample. These may be either enhanced selectivity of sample ionization (electron capture detector, surface ionization detector) or additional chemical–structural information about a component of a sample (mass selective detector). 1–3 Though a remarkably high level of column efficiency can be obtained with commercially available bonded-phase capillary columns, the widespread utilization of mass spectrometers as a GC detector in environmental, industrial, and medical measurements is evidence that addi- tional dimensions of information in GC separations have importance. Thus, plots of ion abundance versus m/z orthogonal to retention time often add analytical confidence sufficient to justify the high costs of purchase, operation, and maintenance of mass spectrometers as GC detectors. In 1982, a conventional IMS drift tube design was described for use as a GC detector for capillary columns, where sample clearance in the source region was fast and sample neutrals were prevented from diffusing into the drift region. 4 This detector exhibited high-speed response, low memory effects and repro- ducible gas phase ion chemistry inside the drift tube. Recently, class specific information has been discovered in mobility spectra under certain conditions of low moisture and high temperature. 5,6 Consequently, mobility spectrometers can be adjusted to provide spectra either with intact product ions or with fragment ions so that information density might be controlled by the analyst. In such conditions, confidence levels in categorizing mobility spectra by chemical class may be very high (for chemicals class 90–95%, for chemical identification inside classes ~ 80%). In summary, the information density in mobility spectra is sufficiently high so that mobility spectrome- ters may be considered economical and sensible alternatives to a mass spectrometer as GC detectors when utilities, size, weight and cost are restrictions. One example is air quality monitoring on-board the international space station where measurements are now made using a GC/IMS instrument, the Volatile Organic Analyzer. 7 A significant limitation to the widespread use of the conventional mobility spectrometers as GC detectors has been the cost of manufacturing drift tubes. Compared to mass spectrometers, mobility spectrometers are simple and inex- pensive; however, traditional mobility detectors can be con- sidered costly compared with the FID or a thermal conductivity detector. Though miniaturization of drift tubes may be helpful in reducing costs of manufacture of mobility spectrometers, 8 the best route to low cost detectors could be a combination between miniaturization and methods of mass production. A drift tube that may be regarded as economical has been described and is a planar micro-machined drift tube 9–11 with rectangular dimen- sions of 25 mm long 3 0.5 mm deep 3 5 mm wide. This drift tube was operated using a non-traditional approach with high field asymmetric waveform dependent mobility methods. 12,13 Though the drift tube is small and ion losses to the walls might be considered a complication, sub-nanogram detection limits were obtained in the absence of ion shutters. Since these drift tubes are manufactured using mass fabrication methods, the most expensive part of the IMS analyzers, the drift tubes, now can be made in mass with concomitant low costs. The size of the analyzer and supporting electronics suggest that this design may be attractive especially for field gas chromatographs and process analyzers. In a high field asymmetric waveform-operated drift tube, ions are transported through the drift tube by a gas flow and electric fields are applied perpendicular to the ion transport through a planar drift region. 12,13 Unlike the behavior of ions in traditional drift tubes, ion separation in this drift tube occurs according to the ion mobility dependence on the electric field intensity. The small size of the miniature mobility analyzer allows ion residence times of about 1 ms and consequently the scan time This journal is © The Royal Society of Chemistry 2002 466 Analyst, 2002, 127, 466–471 DOI: 10.1039/b111547m