Adsorption of Trinitrotoluene on Uncoated Silicon Microcantilever Surfaces L. A. Pinnaduwage,* D. Yi, F. Tian, and T. Thundat Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, Tennessee 37831-6122, and Department of Physics, University of Tennessee, Knoxville, Tennessee 37996 R. T. Lareau Department of Homeland Security, Atlantic City, New Jersey 08405 Received September 5, 2003. In Final Form: January 13, 2004 We measured the adsorption characteristics of trinitrotoluene (TNT) on piezoresistive silicon micro- cantilever surfaces under ambient air using a well-characterized TNT vapor generator. This allowed us to quantify the adsorption parameters and to estimate the sticking coefficient. The sticking coefficient initially increases with TNT exposure time and then levels off around 0.3. Atomic force microscopy images of silicon surfaces exposed to TNT revealed “island” formation of the adsorbate on the silicon surface. At low exposure times, mainly the number density of islands increased with exposure time; at longer exposure times, the size (in particular, height) of the islands grew, corresponding to the higher sticking coefficients. These observations can be qualitatively explained via the difference between TNT-surface and TNT-TNT interactions mediated by water molecules. 1. Introduction Adsorption and desorption of explosive material on various surfaces is of importance in tracing concealed explosives. 1 On the other hand, explosive vapor detection can be hindered by the tendency of these vapors to be adsorbed on the surfaces of delivery tubes. 2 Therefore, understanding adsorption/desorption characteristics for various surfaces is of importance for explosive detection. A third case involves the adsorption/desorption charac- teristics of the explosive sensor surface itself. Currently we are developing an explosive vapor detection scheme in which explosive vapors are deposited on a piezoresistive cantilever and are heated to temperatures above their deflagration points by applying a 10 V pulse of millisecond duration. 3,4 Since silicon-based microelectromechanical systems (MEMS) sensors such as microcantilevers offer simple and inexpensive solutions for explosive vapor detection, 5 a detailed understanding of the adsorption and desorption characteristics of explosive vapors on silicon surfaces is important. To our knowledge, there have been only four studies conducted up to now on desorption properties of trini- trotoluene (TNT) on silicon surfaces, 6-9 and no quantita- tive studies have been conducted on the adsorption of TNT. In the studies conducted by Mu et al., 6,7 the surface was cooled to liquid nitrogen temperature of -195 °C and the TNT source was heated to 93 °C; during the course of deposition, the entire system was maintained under helium gas purge to prevent possible water condensa- tion. 6,7 Under these conditions, it was reported that an amorphous film of TNT was deposited on the silica surface. 6 We have studied desorption characteristics of TNT and other species relevant for microcantilever-based detection of explosives recently; 8,9 those studies were conducted with vapor sources that were not characterized; that is, neither the temperature of the source nor the flow rate of the vapor stream was controlled. In the present study, we used a well-characterized TNT source that allowed us to control the amount of TNT delivered to the microcantilever and thus enabled us to quantify the adsorption param- eters. Furthermore, these studies were conducted under ambient conditions where water vapor plays a significant role, as discussed below. 2. Experimental Method The experimental apparatus used in the present experiments is shown in Figure 1. The TNT vapor was generated by a vapor generator developed at Idaho National Engineering and Envi- ronmental Laboratory (INEEL). The vapor stream was generated by flowing dry air through a reservoir containing TNT. The reservoir consisted of 0.1 g of TNT deposited on glass wool contained in a stainless steel block. The reservoir temperature was controlled via thermoelectric elements that cooled or heated the reservoir, generating a given level of vapor saturation within the reservoir. The vapor generated in the reservoir is delivered through a delivery tube as shown in Figure 1. This stainless steel delivery tube has a 2 μm perforated inner tube. The carrier gas (dry air or N2) flows continuously through the clear hole in the stainless steel block and enters the outer tube, goes through the perforated inner tube, and merges with the flow coming from the reservoir during a “TNT pulse”. The purpose of this “tip” air stream is to prevent the loss of TNT in the delivery tube. For the data presented in this paper, the reservoir flow was 200 standard (1) Bender, E.; Hogan, A.; Leggett, D.; Miskolczy, G.; MacDonald, S. J. Forensic Sci. 1992, 37, 1673-1678. (2) Peterson, P. K. Proceedings of the International Symposium on the Analysis and Detection of Explosives, Mar 29-31, 1983; U.S. Dept. of Justice: Washington, DC, 1984; pp 391-395. (3) Pinnaduwage, L. A.; Gehl, A.; Hedden, D. L.; Muralidharan, G.; Thundat, T.; Lareau, R. T.; Sulchek, T.; Manning, L.; Rogers, B.; Jones, M.; Adams, J. D. Nature 2003, 425, 474. (4) Yinon, J. Anal. Chem. 2003, 75, 99A-105A. (5) Pinnaduwage, L. A.; Boiadjiev, V.; Hawk, J. E.; Thundat, T. Appl. Phys. Lett. 2003, 83, 1471-1473. (6) Mu, R.; Ueda, A.; Wu, M. H.; Tung, Y. S.; Henderson, D. O.; Chamberlain, R. T.; Curby, W.; Mercado, A. J. Phys. Chem. B 2000, 104, 105-109. (7) Mu, R.; Ueda, A.; Liu, Y. C.; Wu, M. H.; Lareau, R. T.; Henderson, D. O. Surf. Sci. 2003, 530, L293-L296. (8) Muralidharan, G.; Wig, A.; Pinnaduwage, L. A.; Hedden, D.; Thundat, T.; Lareau, R. T. Ultramicroscopy 2003, 97, 433-439. (9) Pinnaduwage, L. A.; Thundat, T.; Gehl, A.; Wilson, S. D.; Hedden, D. L.; Lareau, R. T. Ultramicroscopy, in press. 2690 Langmuir 2004, 20, 2690-2694 10.1021/la035658f CCC: $27.50 © 2004 American Chemical Society Published on Web 02/25/2004