At present, chemical and biomedical sensing techniques rely upon analyses performed by a variety of instruments using a number of different sensing technologies. Most of these sensing technologies are expensive and suffer from intrinsic limitations in volume, weight, and power that impede their practical implementation as sensors. Current technologies, therefore, are not well-suited to integration into a single device or adaptation as a basis for the development of a universal platform for multiplexed detection of chemical and biological species. In addition, at present, most bioassays use labels such as fluorescent tags for identification, increasing the number of steps involved in detection. These limits may be alleviated soon with the advent of micromechanical sensor arrays that exploit the nanomechanical forces induced by molecular adsorption as a basis for sensing.1,2
Microcantilever sensors can measure the extremely small forces caused by molecular adsorption. The approach offers an exciting opportunity for the development of highly sensitive chemical and biological sensors based these forces. Recent advances in microlithographic technologies and microfabrication techniques allow us to use batch-processing techniques to manufacture the sensors on silicon wafers and other materials in a cost-effective and modular fashion. In fact, these attributes constitute key enabling factors for the technology.
Cantilever beams undergo bending due to surface stresses created by molecular adsorption when that adsorption is confined to a single side of the cantilever. A cantilever resonance frequency also varies as a function of molecular adsorption caused by mass loading. We can measure adsorption-induced bending and frequency variation using techniques such as variations in optical beam deflection, piezoresistivity, piezoelectricity, capacitance, and electron tunneling. Let's focus in on the piezoresistive method, which offers compatibility with miniaturization, fabrication into arrays, the ability to work in air as well as under solution, and high sensitivity.3
Figure 1: A 1 X 4 array of peizoresistive cantilevers contains four elements, each 120 µm long and separated by 470 µm.
The piezoresistive method exploits bending-induced changes in the resistance of a silicon cantilever. For the cantilever to have any observable piezoresistivity, the electrical conductivity along its thickness has to be asymmetric, which is often accomplished by differential doping. We can observe piezoresistivity in a silicon cantilever coated on both sides with silicon nitride (SiN), if we apply the SiN coating asymmetrically so that the neutral axis of the cantilever is inside the coating (see figure 1). Our experiments with piezoresistive cantilevers show that transducers based on this technology feature excellent sensitivity with noise levels below 1 nm.
Although cantilever sensors are extremely sensitive, they offer no intrinsic chemical selectivity. Cantilever sensors can achieve chemical selectivity if we coat them with chemically selective thin films, such as polymers, self-assembled monolayers, antibodies, or peptides. Microcantilevers can thus essentially become universal platforms for measuring a multitude of chemicals, biomolecules, and even organisms, depending on the type of coating. Sensing the Environment
Our group has used cantilevers to detect chemical species such as chemical warfare agents, volatile organic compounds, and toxic industrial compounds. The devices have also registered a number of bio-molecules, organisms, and bio-chemicals, including markers for cancer and cardiac disease, DNA markers, bio-toxins, bio-warfare pathogens, glucose, and calcium ions.
Figure 2: Piesoresistive cantilever deflection (as a function of exposure time to a stream of RDX vapor in dry nitrogen) shows the cantilever bends as a result of exposure to the vapor and returns to its original position without any observable hysteresis when exposure ends; the RDX concentration was 6 ppt.
As a demonstration of vapor-phase detection, we carried out experiments using explosive vapors of cyclotrimethylene trinitramine (RDX). To achieve chemical selectivity, we used two cantilevers in a four-cantilever array with self-assembled monolayers of 4-mercato benzoic acid. The other two cantilevers served as references for common-mode rejection. We then exposed the array to a vapor stream of 6-ppt RDX in dry nitrogen (see figure 2). As the figure shows, the cantilever bent as a result of RDX adsorption and returned to its normal position once the RDX vapor flow stopped. The response to RDX proved to be completely reversible with no appreciable hysteresis. We have shown that the cantilever sensor can respond to RDX in 10 s and regenerate itself within 50 s. Detection of a target compound in the presence of multiple vapors will require the use of larger arrays, many selective or partially selective coatings, and pattern recognition for analysis.
To demonstrate the feasibility of biological applications, we carried out DNA hybridization (binding) by immobilizing thiol-modified 20-base-long single strands of DNA (probes) on two cantilevers in a four-element array. The cantilevers were coated on one side with gold. When we introduced complementary single strands of DNA (targets) to the liquid cell holding the cantilever array, the cantilevers with DNA probes deflected as a result of the binding of the probe strands and the target strands and formed double-stranded DNA. The cantilever deflection resulted from the reduction of compressive forces on the gold side of the cantilever caused by double-stranded DNA formation. Our studies have shown that the extent of cantilever bending varies as a function of the length of the complementary single-strand DNA. Applications and Challenges
Simplicity, low power consumption, cost-effectiveness, inherent compatibility with array designs, and the ability to operate in air or liquid make cantilever sensors very attractive for a variety of applications. Piezoresistive signal transduction is well-suited to integration with on-chip electronic circuitry. Use of array and pattern recognition will obviate the need for coatings that individually respond to only one particular chemical. Currently available micromachining technologies could yield multiple target sensor arrays involving tens of cantilevers, analog processing, and even local telemetry, all on a single chip. We can lower noise, increase selectivity, and enhance robustness by increasing the number of sensing elements in an array.
A number of challenges must be overcome before cantilever array sensors can come into widespread use. The technology for designing electronic chips is advanced, but the integration of electronic, mechanical, and fluidic designs still needs work; efforts are underway to accelerate the design of fully integrated devices. More advances are needed in the design of chemically selective layers and their efficient immobilization on cantilever arrays. The development of on-chip microfluidics will also be needed to increase response time, especially in biological applications.
Potential applications are widespread within the consumer, military, industrial, and clinical markets. This will be especially true when mixtures of multiple target molecules can be analyzed and screened on a single, miniature chip. Telemetry will enable the use of mobile units worn or carried by personnel, facilitate the deployment of fieldable devices to relay pertinent data to central collection stations, and may even replace wired sensors in some applications. oe
- T. Thundat, P. Oden, and R. Warmack, Microscale Thermophysical Engineering, 1;185 (1997).
- T. Thundat and A. Majumdar, "Microcantilevers for physical, chemical, and biological sensing," Sensors and Sensing in Biology and Engineering, Barth, Humphry, Secomb (eds.) p. 338, Springer Wein, New York (2003).
Thomas Thudat is a distinguished scientist at Oak Ridge National Laboratory, Oak Ridge, TN.