Latest News from NevadaNano

J. D. Adams1,a,b, B. Rogers a,b, R. Whittena
Nevada Nanotech Systems, Inc., 661 Sierra Rose Drive, Reno, NV USA 89511
Department of Mechanical Engineering, University of Nevada, Reno, Mail Stop 312, Reno, NV USA 89557


The remarkable sensitivity, compactness, low cost, low power-consumption, scalability, and versatility of microcantilever sensors make this technology among the most promising solutions for detection of chemical and biological agents, as well as explosives. The University of Nevada, Reno, and Nevada Nanotech Systems, Inc (NNTS) are currently developing a microcantilever-based detection system that will measure trace concentrations of explosives, toxic chemicals, and biological agents in air. A baseline sensor unit design that includes the sensor array, electronics, power supply and air handling has been created and preliminary demonstrations of the microcantilever platform have been conducted. The envisioned device would measure about two cubic inches, run on a small watch battery and cost a few hundred dollars. The device could be operated by untrained law enforcement personnel. Microcantilever-based devices could be used to “sniff out” illegal and/or hazardous chemical and biological agents in high traffic public areas, or be packaged as a compact, low-power system used to monitor cargo in shipping containers. Among the best detectors for such applications at present is the dog, an animal which is expensive, requires significant training and can only be made to work for limited time periods. The public is already accustomed to explosives and metal detection systems in airports and other public venues, making the integration of the proposed device into such security protocols straightforward.

Keywords: microcantilever, chemical, biological, explosive, sensor


Maritime shipping of containers is a major part of world commerce. Any terrorist activity that might utilize the container shipping process would likely disrupt a large portion of the shipping industry and spark a worldwide recession, if not depression.1 Yet in spite of the threat to the worldwide economy, the ability to detect unlawful and hazardous materials in a container is very limited. Recently the Container Security Initiative and the Customs-Trade Partnership Against Terrorism have made strides to improve container security, including placing custom agents in foreign ports to monitor security procedures and inspect high-risk containers before shipment to the U.S. Unfortunately these agents lack independent, hard data on the contents of the containers. Our objective as a company was to help solve this problem and provide agents with data taken from air samples inside the container measuring the concentrations of unlawful, lethal and hazardous material in the containers.

We have designed, modeled, built and tested all components of a prototype sensor system for measuring the concentrations of unlawful or hazardous materials. The main tasks were building and testing a prototype detector against several potential threat agents. There are three categories of material this system is designed to detect—chemical, biological and explosive. To test the effectiveness of the system, we tested seven chemical agents—including toxic industrial chemicals ammonium hydroxide, toluene diisocyanate (2,4), formaldehyde, and allyl alcohol. We also used microcantilever sensors to identify Serratia marcescens, a biological agent analog to the plague, from among other bacteria using a technique not demonstrated before using cantilevers to our knowledge. Finally, we detected two explosive vapors: trinitrotoluene (TNT) and pentaerythritol tetranitrate (PETN). With the completion of these demonstrations, we find this technology to be very well suited for advanced security applications.

Figure 1: Depiction of an embedded, secure, wireless sensor unit welded on the inside of shipping container with an external antenna.

Figure 2: Depiction of shipping container with embedded sensor at different locations in the shipping cycle.

The nature of our detection methodology relies on detecting vapor from hazardous material in the air. This method benefits from extended outgassing time of the hazardous materials, so the detection probability improves over time. As a result, continued monitoring of the air in the container during the entire shipping process will provide the highest security. We also recognize that the cost and threat of hazardous material in the container can be minimized if the material is detected at the earliest possible time—ideally prior to arrival at the initial shipping port.


2.1 Determine chemicals, biological agents and explosives of interest

In determining which substances pose the greatest threat to homeland security we have researched high explosives, toxic chemicals, and biowarfare agents—gaining input from the shipping industry and from experts in the Department of Homeland Security and the Customs and Border Patrol. While we would eventually provide a comprehensive solution for all chemical, explosive and biological threats, we have attempted to prioritize these agents to maximize the threat detection capabilities of the initial system. For the purposes of this proof-of-principle work, all lethal, highly toxic, or explosive substances are good candidates for our sensing system.

We chose the USACHPPM list of toxic industrial chemicals2 as our baseline list, and chose the following lethal agents for testing: Toluene diisocyanate (2,4), Formaldehyde, Ammonium Hydroxide, Allyl Alcohol. The toxicity levels for these chemicals ranges from 0.51 ppm/hr for toluene diisocyanate (2,4) to 1100 ppm/hr for ammonium hydroxide. We also conducted tests of other chemical vapors, including toluene, acetone and ethanol. The ability to distinguish between different types of chemicals and in particular to accurately distinguish between lethal and non-lethal, legal and illegal, and expected and unexpected chemicals is an important design criteria of the system. The additional measurements of the 3 common chemicals—toluene, acetone and ethanol—provided additional evidence of the sensor array’s ability to distinguish between a variety of chemicals.

For explosive detection, we chose to demonstrate detection of TNT and PETN based on that fact that these are among the most common explosives3 (see Table 1). They are also representatives of high vapor pressure (TNT) and low vapor pressure (PETN) explosives. Low vapor pressure explosives can be harder to detect.

Table 1: List of widely used explosives and their major ingredients – PETN, RDX and TNT being the most common (taken from reference [3]).

For biological agent detection, there are numerous biological agents to target. The three primary classifications of biological warfare agents, the key features of each, and examples of each type are available from the U.S. Centers for Disease Control and Prevention Category A threats.4 Category A is the highest level of threat. Agents in this category “can be easily disseminated or transmitted from person to person, result in high mortality rates, and have the potential for major public health impact.” As with the selection of Toxic chemicals for the system demonstration, we used the CDC table for selecting a target bio threat.

2.2 Cantilever Array Design

Selectivity is an important design parameter for a useful sensor system. Cantilevers can be easily fabricated in arrays utilizing batch production techniques common to semiconductor factories. The use of an array of cantilevers allows us to use several cantilever variations and numerous polymer coatings in the array to obtain sensitivity to various analytes simultaneously.

Cantilever arrays have been assembled to test their capabilities for detecting hazardous materials. This design utilizes individual cantilevers. The cantilever used is the Active Probe, a piezoelectric atomic force microscopy cantilever available from Veeco, Inc. The integrated piezoelectric film enables these cantilevers to be self-sensing: no external optical system is needed to monitor the beam motion. The cantilevers are arranged into an array inside customized sample chambers like the Plexiglas unit shown in Figure 3.

Figure 3: On the left, a cantilever sensor chamber for chemical vapor sensing tests. Chemical vapor enters the chamber through the inlet, passes over the cantilevers mounted inside the chamber (this photo shows 7 cantilever substrates and their corresponding wire pairs), and out the other side of the chamber. On the right, a scanning electron micrograph of the Active Probe, showing its piezoelectric sensor and actuator. This cantilever is commercially available as an atomic force microscope probe tip and has not been optimized for chemical sensing applications.

For chemical vapor detection experiments, the array consisted of 9 cantilevers. They are driven by a 1V AC signal, the frequency of which is swept from 35-155 kHz. All of the cantilevers have been coated to varying thicknesses with 9 different polymers. These polymers were dissolved in a solvent (usually chloroform), then spray- or dip-coated onto the cantilevers to a thickness of at least 500 nm on at least one side.

The coating process affects each cantilever’s resonance frequency and serves to spread the resonance peaks across the frequency spectrum. During chemical vapor experiments, the resonance frequency of each cantilever in the array changes with the adsorption of vapor. This is typically a downward shift in resonance frequency, due to the added mass of the vapor, but can also be upward, due to chemical processes affecting the stiffness of the polymer coating. These resonance shifts represent the sensor response.

To address the piezoelectric cantilever array, an input signal sweeps through the resonance frequencies of each of the cantilevers while an RMS converter tracks the amplitude of the sensor output.


Experiments conducted using an array of cantilevers selectively coated with polymers demonstrate both the sensitivity and the selectivity of this sensor platform. Sensitivity is defined here as the minimum detectable concentration of an analyte vapor. Selectivity encompasses the ability of a sensor to correctly identify specific analytes and distinguish one analyte from another.

3.1 Selectivity

The cantilevers’ respective responses to analyte vapors at varying concentrations can be evaluated by pattern recognition algorithms in order to differentiate analyte vapors from one another. Shown in Table 2 are respective data used in principle component analysis (PCA). These data represent the shift in resonance frequency of each cantilever in an array after separate exposures to five chemical vapors. The advantage of using different coatings on different cantilevers in an array for reference and selectivity is gained by using pattern recognition algorithms, like PCA.

Table 2: The shift in resonance frequency for each cantilever in an array when exposed to 10% saturated vapor in air for 10 minutes for each analyte (with positive shifts indicating a decrease in cantilever resonance frequency). Cantilevers are identified on the left by their respective polymer coating (PECH, PDMS, etc.).

The data listed in Table 2 can be graphically represented in a radar plot like that shown in Figure 4. Data of this type can then be used as the input for principle component analysis algorithms used to more clearly identify and differentiate each analyte vapor.

Figure 4: A radar plot using data from Table 2 shows the resonance frequency shift all nine cantilevers in the array to chemical vapors, including toxic industrial chemicals. Each cantilever is coated in a different polymer (PECH, PDMS, etc.). For this experiment, the concentration of each vapor to which the cantilever array was exposed was 10% saturated vapor in air, by volume, for 10 minutes.

The PCA plot in Figure 5 demonstrates the selectivity of this sensor platform. Data depicted in Figure 5 represent sensor responses to 1% saturated analyte vapor by volume in air—as produced by a vapor generation system we designed for such tests. There are two data points for each analyte representing the first and second measurement – taken at 5 and 10 minutes of exposure, respectively. The second measurement should therefore have a larger response due to the increased time in the presence of the analyte. Data were also gathered at 0.1% saturated vapor and indicate that the sensor array was also able to distinguish between all of the analytes at lower concentrations than we expected for this model of microcantilever. Note that the saturated vapor concentrations of 0.1%, 1% and 10% represent different parts per X vapor concentrations depending on the analyte, as shown in Table 3.

Table 3: The concentration of various chemicals at 0.1, 1 and 10 percent saturated vapor, by volume, in air. Concentrations are determined using the ideal gas law.

Figure 5: Principal component analysis of data from an array of coated cantilevers demonstrates selectivity to chemical vapors at 1% saturated vapor by volume. Two dots of each color correspond to sensor response after 5 and 10 minutes of exposure time.

Figure 5 represents a proof-of-principle demonstration of the efficacy of principle component analysis, thereby resolving one of the major unknowns of this sensor technology. We have enlisted the expertise of pattern recognition specialists in further refining this approach to chemical vapor selectivity. This approach shows much promise, appearing to differentiate among chemical vapors at concentrations as low as 770 ppb to 33 ppb.

3.2 Sensitivity

The design of the cantilever sensors in the array determines the ultimate sensitivity of the system. Data thus far have been gathered using the AFM-type cantilever previously described. Figure 6 shows the response to allyl alcohol vapor of a cantilever coated in poly(dimethylaminoethyl methacrylate). As the cantilever’s polymer coating absorbs allyl alcohol vapor, the effective mass of the cantilever increases, which in turn causes the resonance frequency to shift downward. This shift is the sensor response. All tests were conducted at room temperature and pressure. This figure demonstrates the low parts-per-million sensitivity of the detector to a toxic industrial chemical. The resonance shift of the cantilever at approximately 21 ppm is 87 Hz. Based on this data point and other system information, the lowest detectable concentration of allyl alcohol for this proof-of-concept system is below 10 ppm. This value is in agreement with our models for the sensitivity limits for this unoptimized cantilever design. We expect that in the future, system modifications could enable sensitivity limits from parts-per-quadrillion to parts-per-quintillion. Competing sensor technologies5 are not to our knowledge capable of parts per quintillion sensitivities, and technologies which can detect down to parts per quadrillion are traditionally quite large in size.

Sensitivity is a function not only of the cantilever’s geometry (which governs its mass resolution), but also of the polymer coating, and the amount of coating used. Optimum coating thicknesses were not investigated as part of this work but will be a part of future development work.

Figure 6: The response of a cantilever sensor coated in poly(dimethylaminoethyl methacrylate) to allyl alcohol vapor at varying concentrations: 0.1%, 1% and 10% allyl alcohol vapor by volume. The concentrations in parts-per-million shown on the graph are estimates. Allyl alcohol is a toxic industrial chemical.

3.3 Test for Explosives

The microcantilever used to detect explosives contains an integrated piezoresistive heating element. The microcantilever’s dimensions are approximately 300 microns long, 100 microns wide and 6 microns thick—very similar to the Active Probe described earlier. The heating element allows rapid heating of the cantilever to evaluate the thermal response of adsorbed material. The presence of the heating element does not interfere with the ability of the cantilever to accurately measure the mass of the adsorbed material by means of the shift in the cantilever’s resonance frequency. As a result, the cantilever maintains the ability measure mass and thermal properties simultaneously.

Microcantilever sensor platforms have previously been demonstrated to verify the presence of trace explosives in two ways: (1) using uncoated but heater-equipped cantilevers to deflagrate explosives,6 and (2) using a selective 4-MBA2 self-assembled-monolayer coating to detect bending and resonance frequency changes.7 We first tested the deflagration method in collaboration with Oak Ridge National Laboratory in 2003 [6]. In this method, a piezoresistive heating element in the cantilever is used to heat any substance adsorbed on the cantilever. In the case of an adsorbed explosive substance, the heating causes a quick reaction, apparently a deflagration of the material.3 Using a high-speed video camera, the deflagration event can be seen in Figure 7.

2 4-mercaptobenzoic acid
3 Only a trace amount of material undergoes deflagration and due to the low concentrations of explosive present near the sensor and we have seen no evidence that this test reaction can initiate a larger reaction in the environment. The sensor is not damaged by detonation of such minute quantities of explosive.

Figure 7: This three-step sequence (bottom to top) shows magnified high- speed video images of TNT on a microcantilever, TNT deflagration on a microcantilever with the corresponding vapor plume, and the vapor plume receding from the microcantilever, respectively. The video images were taken from an actual experimental sequence, and superimposed on higher- resolution SEM images of self-sensing microcantilever explosive sensors for viewing purposes.

We believe the heat-sensing approach to be a powerful tool in a suite of analytical approaches that can be used to detect and discriminate trace amounts of explosive material from other non- hazardous materials using a microcantilever. This approach would therefore be additional verification for detection with coated cantilevers. During this work, we demonstrated proof of concept for explosive verification and selectivity using the deflagration method.

In earlier work, the high speed video images (Figure 7), the specific voltage required to remove material, and the fact that material could be quickly removed all provided evidence of a small deflagration. In additional tests performed during this program, we were able to remove material with the deflagration method, to detect the removal of material using integrated self sensing, and see a difference in thermal signatures between TNT and two test interferents.

The explosive detection methods tested as part of this work were:

  • Detecting the signature thermal response caused by
    quickly heating trace amounts of explosive material
  • Detecting the signature loss of adsorbed explosive mass at different heat pulse temperatures
  • Detecting the unique adsorption/desorption behavior of explosive material

To initiate deflagration, a voltage pulse is applied to the 3000 ohm integrated piezoresistive layer of the cantilever. This pulse heats the cantilever quickly and typically is programmed to last 10 or 20 ms. The zinc-oxide piezoelectric layer of the cantilever, in addition to serving as an actuator and sensor for cantilever motion, is also sensitive to thermal variations, which manifest as changes in the electrical properties of the piezoelectric film. We have been able to use this sensor for proof of principle explosives detection. The detection follows this general protocol:

The baseline resonance frequency of the cantilever in the test shown in Figure 8 was 67,100 Hz. While exposing the sensor to TNT vapor, explosive material adsorbed onto the cantilever and its resonance frequency quickly dropped to 65,600 Hz. This 1,500 Hz decrease equates to roughly 15 nanograms of adsorbed TNT. As the mass adsorption tapers off, a 10V pulse is applied to the heater (minute 7), which causes no change in the resonance frequency. The first 30V pulse, however, is sufficiently hot to cause a 125 Hz increase in resonance frequency, equating to approximately 1.2 ng of removed TNT. The next 30 V pulse does approximately the same. A 20 V pulse causes a smaller shift—28 Hz, removing about 300 picograms of TNT. The temperature achieved by a 20 V pulse, at least in the local region of the integrated heater, appears to be hot enough to remove a small amount of TNT and a 30 V pulse removes more, as expected.

Tests were also conducted using PETN as the explosive vapor analyte. These tests provided similar graphs that showed that adsorption of PETN and subsequent removal of PETN by the thermal pulses.

Figure 8: This experiment shows how TNT mass adsorbs onto the cantilever, lowering its resonance frequency and how this added mass can be removed by successive heating pulses. While 10 V pulses are not hot enough to remove TNT mass, 20 V and 30 V pulses are.

Also, a preliminary control experiment of 30V pulses occurring in the presence of air, explosive vapor and control substances such as plastic vapor, solvents, dust and water vapor indicates that explosive materials have a unique heat signature that can be used to detect explosive vapors in air. These control substances were vaporized in the presence of the cantilever, which subsequently adsorbed mass. As this mass desorbed from the sensor, the rate of desorption was measured, and a different desorption rate was measured for each substance. After the desorption was measured, a series of 30 V heat pulses were applied to the cantilever to remove the adsorbed mass, and the amount of mass removed per pulse was also measured. Both water and the solvent, acetone, did not significantly adsorb, and desorbed quickly before the comparison measurements were made between the other substances. The three substances that did adsorb were TNT, plastic and dust. The dust did not desorb appreciably, while the plastic did and the TNT desorbed even more quickly. These are all consistent with the vapor pressures of the materials.

3.4 Test for Biological Agents

All bacteria, as living organisms, give off unique mixtures of exhaled gases as part of their metabolism. These characteristic “odors” can be detected by sorptive polymer array sensors and differentiated using pattern recognition techniques in order to identify bacterial strains. This approach has been demonstrated for diagnosing bacterial infections by sampling urine vapor,8 to detect pulmonary infections in patents’ breath,9 to predict bacterial type and culture growth phase,10 to classify grains,11 and to detect and identify a range of microorganisms.12,13,14 While the use of microcantilevers as a transduction platform for biological sensors has been well documented, this technique for identifying bacteria with microcantilevers has not been demonstrated to our knowledge.15 We have used this vapor sampling technique to demonstrate detection and differentiation of a biological warfare agent analog. We selected the bacteria, Serratia marcescens, used in prior studies as an analog for the Category A bacterial threat, Yersinia pestis (plague).16 The experiment entailed sampling the headspace vapor of a bacterial cultures using the prototype sensor array. Pattern recognition techniques were used to discriminate the bacteria. Bacterial samples and controls were prepared by Dr. Eric Marchand and his staff at the University of Nevada, Reno.

Figure 9: Principal component analysis of data collected by an array of polymer-coated cantilever sensors when exposed to the headspace vapors of various bacteria and control samples. There are two data points per analyte—taken after 4 and 6 minutes of exposure to headspace vapor. Analytes are: Water vapor control (no bacteria) grown on ATCC Media; Water vapor control (no bacteria) on LB Media; Serratia marcescens (plague analog) on ATCC media; E Coli ATCC on ATCC media; E Coli DH5α on ATCC media; Both Serratia marcescens and E Coli ATCC grown together on ATCC media; E Coli DH5α grown on LB media.

Three separate tests confirmed that the headspace vapor of certain bacteria can be used to discriminate among them. The respective resonance frequencies of each of the cantilevers in the array during exposure to bacterial headspace vapors were subtracted from the cantilevers’ resonances in clean room air and used in principal component analysis, which plotted the resulting principal components of the data on three axes. See Figure 9.


An advanced container security system has been designed to monitor maritime and other shipping containers for the presence of toxic chemicals, explosives and biological warfare agents. A prototype of the sensor has been fabricated and tested using off-the-shelf components. This prototype successfully identified trace concentrations of four Toxic Industrial Chemicals, two explosives and one Class A bioagent (analog). The performance of the system, which currently uses cantilevers not optimized for this application, met or surpassed our sensitivity estimates. Overall, the feasibility of an advanced container monitoring system has been demonstrated. We stand ready to fabricate field suitable prototypes for field testing.


The authors thank the U.S. Department of Energy (DOE) and NA-22, Office of Nonproliferation Research and Engineering, for supporting this work. (Contract# DE-FC52-04NA25657.)


1 “Cargo Security” p74 Technology Review June 2004
3 J. Yinon, “Field detection and monitoring of explosives,” Trends in Analytical Chemistry, vol. 21, no. 4, p. 292 (2002).
6 L. A. Pinnaduwage, A. Gehl, D. L. Hedden, G. Muralidharan, T. Thundat, R.T. Lareau, T. Sulchek, L. Manning, B. Rogers, M. Jones, J.D. Adams, “A microsensor for trinitrotoluene vapour,” Nature, v 425, 474 (2003)
7 L. A. Pinnaduwage, V. Boiadjiev, J. E. Hawk, and T. Thundat, Appl. Phys. Lett., Vol. 83, No. 7, 18 August 2003 – page 1471-1473
8 S. Aathithan, J. Plant, A. Chaudry, G. French, “Diagnosis of Bacteriuria by Detection of Volatile Organic Compounds in Urine Using an Automated Headspace Analyzer with Multiple Conducting Polymer Sensors,” J. Clinical Microbiology 39, No. 7, p. 2590- 2593 (2001).
9 B. Maher, “The E-Nose: Scientists Compete with Nature’s Prolific Sniffers: Electronic sensors explain how being sick can stink,” The Scientist, Vol. 16, issue 19, p. 38 (Sep. 30, 2002)
10 J W Gardner, M Craven, C Dow & E L Hines, “The prediction of bacteria type and culture growth phase by an electronic nose with a multi-layer perceptron network,” Meas. Sci. Technol. 9 (1998) 120–127.
11 J.R. Stetter, M.W. Findlay, Jr., KM. Schroeder, C. Yue and W.R. Penrose, “Quality classification of grain using a sensor array and pattern recognition,” Analytica Chimica Acta, 284 (1993) l-11
12 T.D. Gibson, O. Prosser, J.N. Hulbert, R.W. Marshall, P. Corcoran, P. Lowery, E.A. Ruck-Keene, S. Heron, “Detection and simultaneous identification of microorganisms from headspace samples using an electronic nose,” Sensors and Actuators B 44 (1997) 413–422
13 Martin Holmberg, Fredrik Gustafsson, E. Gunnar Hörnsten, Fredrik Winquist, Lennart E. Nilsson, Lennart Ljung and Ingemar Lundström, “Bacteria classification based on feature extraction from sensor data,” Biotechnology Techniques, Vol 12, No 4 (April 1998) pp. 319–324
14 C.M. McEntegart, W.R. Penrose, S. Strathmann, J.R. Stetter, “Detection and discrimination of coliform bacteria with gas sensor arrays,”Sensors and Actuators B 70 (2000) p. 170–176
15 Álvarez, M., Carrascosa, L.G., Tamayo, J., Calle, A., and Lechuga, L.M. 2003. Nanomechanics for specific biological detection. Proceedings of SPIE Vol. 5118, p. 197-206.
16 Utrup LJ, Frey AH., “Fate of bioterrorism-relevant viruses and bacteria, including spores, aerosolized into an indoor air environment,” Exp Biol Med 4 ( 2004) 345-350.
  Journal Papers