Sensitive microsensors recognize chemical warfare agents.

Researchers at NIST have developed microsensors that detect nmol/mol quantities of chemical warfare agents (CWAs), including sarin (GB), tabun (GA) and sulfur mustard (HD). The recent results represent a significant step forward in the development of low power, CWA detectors for personal protection applications in both military and civilian settings. The project has been supported by the Defense Threat Reduction Agency.

The sensors are silicon-based microelectromechanical (MEMS) devices consisting of suspended 100 [micro]m x 100 [micro]m heatable platforms that have metal oxide sensing films deposited on their surfaces. Gases that adsorb on the surfaces can change the electrical conductivities of the thin, nanostructured films which are typically tin oxide (Sn[O.sup.2]), titanium oxide (Ti[O.sup.2]), zinc oxide (ZnO), or iron oxide (F[e.sup.2][O.sup.3]). The unique features of these sensors are that they can easily be fabricated in arrays and operated in a dynamic temperature mode to generate varied electrical response "fingerprints" for specific gases. In contrast to other conductometric gas sensors, these microsensors can distinguish different analytes and respond much faster (typically in seconds).

Microsensor testing on the CWAs was performed at the Edgewood Chemical Biological Center (ECBC), a Department of Defense facility at Aberdeen Proving Ground, MD. In the tests, four-element arrays were exposed in air backgrounds to individual concentrations of GA, GB, and HD ranging between 4 nmol/mol and 200 nmol/mol. Good signal-to-noise was observed for all of the sensing films employed in these studies (Sn[O.sub.2] and Ti[O.sub.2]), even at the lowest concentrations presented to the devices. Matched pairs of each of the two oxides were used, and excellent sensing reproducibility of the twins was observed. Good stability of the devices also was recorded for hours of CWA exposure. In addition, artificial neural network (ANN) signal processing methods were used to extract the most important and reliable analytical information from the complex temperature programmed responses. Initial ANN successes have demonstrated an ability to recognize individual agents and related molecular simulants and predict their concentrations. Future studies will determine the ultimate sensitivities of the devices as alarm triggers and investigate methods that ensure reliable performance even in the presence of interfering gases, such as diesel exhaust fumes.

CONTACT: Steve Semancik, [phone]; stephen. [e-mail].