This is for educational purposes only. No profit is being made.
Nature 30 July 1998
News and Views
Nature 394, 417 - 418 (1998) © Macmillan Publishers Ltd.
A sense for landmines

ANTHONY W. CZARNIK

A heightened sense of smell, achieved by proxy through chemistry, offers solutions to problems where biologically based sensors are inappropriate. Detection of the TNT in landmines using fluorescent chemosensors is one example.

If the human sense of smell were as keen as a dog's, our perception of the world would be very different. A child could sense its mother in another room. A border agent would know when drugs were being smuggled in the car passing by. And a soldier clearing a field of landmines could avoid stepping or kneeling on the one buried just in front of him (Fig. 1). Of course, most dogs' sense of smell is much better than ours, and there's no immediate prospect that this biological fact will be undone. But non-biological enhancement of our sense for odours can be achieved with sensors -- devices, molecule-sized or larger, that signal the presence of matter (or energy) in a way discernible by one of the human senses.

Figure 1 Bosnia, 1995 -- United Nations troops sift through snow, ice and mud in a search for landmines.   
Writing in the Journal of the American Chemical Society1, Yang and Swager describe the evaluation of porous, shape-persistent, fluorescent polymer films that serve as chemosensors for explosives -- in particular, trinitrotoluene (TNT) and its derivatives, which are commonly used in landmines. Sensors for TNT are in routine use at airports, for instance, but the technology requires sophisticated instruments and simply is not rugged enough for field use. And other ways of detecting landmines are far from ideal -- dogs tire easily and, with their requirement for a full-time trainer, are expensive; ground-penetrating radar and metal detectors lack the specificity of chemical sensors and produce too many false positives.

Chemosensors -- sensors incorporating a receptor element that is not derived from biology2,3 -- have enjoyed increasing attention as alternatives to biosensors. Occasionally, the alternative is required because no biological receptor has been discovered for a given analyte (the chemical species one wishes to detect). More often, the alternative is needed because the biological receptor does not have the properties that are ultimately useful. Although biological macromolecules such as antibodies have deliciously high affinity for many analytes, they are not very rugged molecules. Shake most proteins too hard, and they lose activity. Heat most proteins to too high a temperature, and they denature. Eat most proteins, and they are hydrolysed to amino acids.

For the soldier in the field running mine-clearance programmes, such delicacy is a real problem. Refrigeration is rarely available, and equipment inevitably gets bashed about. In the case of landmine detection, the chance of a false negative stemming from 'dead' antibody in the sensor presents an intolerable risk. So although antibodies to TNT exist, this biological receptor has not yet been used in a field device. Instrumental detection methods, such as mass spectrometry, are ultra-sensitive, and there have been impressive advances in portability. But they suffer from the same lack of in-the-field ruggedness.

If it is ultimately impossible to make antibodies that are robust enough, then the search for receptors must begin with ruggedness. Synthetic receptors, which need not rely on a metastable three-dimensional structure or hydrolysable backbone, fit the bill. A pH indicator strip incorporates a 'receptor' for protons; that receptor is very rugged, and bottled strips can sit on the shelf for years yet retain their activity. The study of synthetic receptors is not new; indeed, there is an extensive literature of non-applied research on the topic. The literature on work that has been 'reduced to practice', by contrast, is much more limited (examples include the stabilization of pharmaceutical products through the inclusion of cyclodextrin and the sensing of calcium-ion concentrations in living cells).

In the sensor business, 'reduced to practice' means, 'meets all the demands of the application' -- no more, no less. For TNT sensing above buried landmines, those demands are predictable:

Sensitivity (that is, transmission of a stable signal resulting from the sensing of concentrations of airborne TNT of just a few parts per billion).

Selectivity (discrimination of those low concentrations of TNT from much higher concentrations of other airborne chemical species).

Real-time signal development (the signal appears within a few seconds, even in the cold; faster is better).

Range (the sensor does not become saturated at lower than the usual concentration of analyte).

Reversibility (remove the TNT and the sensor returns to its initial state within a few seconds; again, faster is better).

Stability (tolerant of water; but active for months in the solid state at high temperatures, such as those typical of deserts).

Yang and Swager's work1 has not yet been 'reduced to practice', but it is a necessary and innovative step towards that goal. Swager's laboratory previously showed that conjugated, polymer-based fluorescent chemosensors can provide signal amplification for the innately sensitive fluorimetric method4, which involves the synthesis of luminescent polymers that are quenched by analyte molecules along the entire length of the polymer. However, such films of flat molecules suffered from 'pi-stacking' and the fluorescence self-quenching that accompanies it. Yang and Swager have now made films of polymers derived from fluorescent pentiptycene molecules that are not flat. As predicted, this film showed less quenching than the first film (although the two systems are not perfectly comparable). But quenching does occur when the flat TNT molecule is introduced. The authors postulate that the three-dimensional pentiptycene polymer has pores, into which the flat TNT molecule can enter and quench fluorescence. This quenching results in a decrease in fluorescence intensity, which serves as the physical basis for signal transduction in the presence of TNT (Fig. 2, overleaf). Several other flat organic molecules, such as benzophenone and 1,4-dicyanobenzene, did not induce fluorescence quenching when tested -- that is, the sensor film can discriminate between TNT and other molecules of similar size and electronic character.

Figure 2  Chemistry of the prototype TNT chemosensor devised by Yang and Swager1 (top), with the geometrical relationships being depicted below. The scheme uses a fluorescent polymer film, and because the pentiptycene molecules are not all flat, the film contains pores into which TNT molecules can enter. Binding of a TNT molecule to the polymer results in  quenching of fluorescence -- the non-fluorescent complex shown here. The consequent change in intensity can be easily measured, and serves as a signal indicating the presence of TNT. As iscussed in the text, this approach has yet to be 'reduced to practice'
To what extent does Yang and Swager's scheme meet the demands necessary for a field TNT sensor? The reported signal amplitude of times20, real-time signal development and reversibility (through rinsing) are all encouraging. The sensitivity to TNT is not yet described, however. And although the discrimination against several planar aromatic compounds is good news, field use may reveal interference by other, as-yet-unknown, compounds; one emerging approach to such challenges of selectivity is the sensor-array format5-7, by which large collections of nonspecific chemosensors all interact with a sample, and the analyte concentration is extracted by network analysis.

None of these drawbacks is cause for pessimism. Research on chemosensors is just starting. Each new discovery adds to the toolbox that will help all of us smell better -- and in this case we can hope, in due course, for a viable field detector for TNT.

Anthony W. Czarnik is at Illumina Inc., 2187 Newcastle Avenue, STE 101, Cardiff, California 92007, USA.
e-mail: aczarnik@illumina.com

------------------

References

  • Yang, J.-S. & Swager, T. M. J. Am. Chem. Soc. 120, 5321-5322 (1998). Links
  • Yoon, J.-Y. & Czarnik, A. W. J. Am. Chem. Soc. 114, 5874-5875 (1992). Links
  • Czarnik, A. W. Accts Chem. Res. 27, 302-308 (1994). Links
  • Zhou, Q. & Swager, T. M. J. Am. Chem. Soc. 117, 7017-7018; 12593-12602 (1995). Links
  • Michael, K. L., Taylor, L. C., Schultz, S. L. & Walt, D. R. Analyt. Chem. 70, 1242-1248 (1998). Links
  • Lavigne, J. J. et al. J. Am. Chem. Soc. 120, 6429-6430 (1998). Links
  • Doleman, B. J., Sanner, R. D., Severin, E. J., Grubbs, R. H. & Lewis, N. S. Analyt. Chem. 70, 2560-2564 (1998). Links


  • Nature © Macmillan Publishers Ltd 1998 Registered No. 785998 England.
    Back to Science & Society