AVS 50th International Symposium
    Homeland Security Topical Conference Wednesday Sessions
       Session HS+MM-WeA

Paper HS+MM-WeA1
MEMS Chemical Sensors for Homeland Security

Wednesday, November 5, 2003, 2:00 pm, Room 309

Session: Detection of Explosives and Other Chemicals for Homeland Security
Presenter: D.C. Meier, National Institute of Standards and Technology
Authors: D.C. Meier, National Institute of Standards and Technology
C.J. Taylor, Pomona College
R.E. Cavicchi, National Institute of Standards and Technology
Z. Boger, National Institute of Standards and Technology
S. Semancik, National Institute of Standards and Technology
Correspondent: Click to Email
Chemical sensors capable of accurate detection of trace quantities of chemical warfare (CW) agents would provide a potent tool for perimeter security, treaty verification, and battlefield threat detection. Ideally, such sensors would be highly sensitive, selective, compact, and require low power. Since many CW agents are lethal at µmol/mol (ppm) concentrations, reliable trace detection is the critical prerequisite to the employment of life-saving countermeasures. In order to meet these goals, microsensors consisting of CVD TiO@sub 2@ and SnO@sub 2@ sensing films on MEMS array platforms have been fabricated. Their response measurements to the CW agents GA (tabun), GB (sarin), and HD (sulfur mustard) in dry air at concentrations between 5 and 200 nmol/mol (ppb), as well as to CW agent simulants CES (chloroethyl ethyl sulfide) and DFP (diisopropyl fluorophosphate) between 250 and 3000 ppb are reported. The devices yield both excellent signal-to-noise and good reproducibility for similar devices. Detection in backgrounds of common battlefield interferants is also discussed. The temperature of each sensor element is independently controlled by embedded microhotplate structures that drive both the CVD process (375 °C) and sensor operation at elevated temperatures (325 to 475 °C). The concentration-dependent analyte response magnitude is sensitive to sensing film growth conditions. Long-term stability studies verify stable sensor responses to GB and HD for 14 hours of agent exposure. Use of fast (200 ms) temperature programmed sensing (TPS) over a broad range (20 to 480 °C) enhances analyte selectivity, since the resulting signal trace patterns include only kinetic data that are unique for each agent tested. Recursive training of an artificial neural network (ANN) assembles the set of most relevant inputs from the TPS data; subsequent validation of the trained ANN verifies positive agent and simulant identification and quantification.