Postal service mulls bioparticle detector for anthrax fight

Two months before the first prototype is due, say the sensor's developers, they are already working with the U.S. Postal Service on a unit specifically designed to sniff out danger in the nation's mail-handling facilities.

Developed with a $1.15 million grant from the Office of Naval Research, the real-time sensor identifies airborne bacteria by shining laser pulses into particles and examining their reactions. "If anything biological, above a certain background level, is detected, then an alarm is sounded," said Jeanne Small, a professor at Eastern Washington who led the research. "At that point specialized test kits can be taken out — little dipsticks that allow you to see with a simple color change if, for instance, anthrax is present." Results come back in under a half hour.

Her device uses an off-the-shelf air sampler from InnovaTek Inc. (Richland, Wash.). The sensor is now being integrated with the air sampler and fitted into the final prototype by Quantum Northwest in Spokane, Wash.

It is scheduled for completion in January.

"When we started the project in 1999, the main emphasis was on biological warfare," said Small. "So we've been tuning up to distinguish bacteria from Army-truck diesel smoke in the presence of bombs exploding — that was the image of what we had to prepare for. But now, of course, things are different. We're thinking about the U.S. Postal Service, letters with anthrax and other biological agents being dispersed throughout our own interior home systems."

Instead of relying on a battery of specific chemical tests to determine whether an airborne material is pathological, Small sought instead their common denominator. She analyzed bacteria tailored for use as weapons, seeking a "universal" set of distinguishing characteristics.

"The more specific a sensor is, the easier it would be for very clever bioterrorists to mess things up just enough that the sensor no longer worked," she said. "In our case, we thought that if we go with something that reflects the total composition of the particle, it would mean that if a terrorist wanted to tweak a protein or two on the surface, that it would not change the overall flavor of how it behaves."

Moreover, the military wanted to detect airborne pathogens with-out using chemicals. Today, airborne bacterial pathogens can be identified only through specific tests: either by culturing the sample or by using a chemical "reagent" tailored to detect an individual pathogen. But that requires the system to know what is being sought before it can be detected. In addition, the chemical reagents need to be continuously replenished.

Chemical-free testing

"The military was looking for something that could be used on the battlefield in remote locations, and the challenge was to detect the presence of bacteria without the use of any chemicals," said Small, a biophysicist. "Reagents are great, they are very specific, but they are costly and you have to keep running the test over and over again. They wanted something that didn't require chemistry, that only used physical mechanisms."

To meet these requirements, Small turned to her expertise in "pulsed-laser photoacoustics," which allows the gross characteristics of airborne bacteria to be inferred from their reaction to a high-energy laser. Essentially, Small cataloged the reactions that bacteria trigger in a laser beam, then used a neural network to differentiate that reaction from that of inanimate airborne particles.

In Small's bioparticle detector, laser pulses excite air samples so that any light-absorbing substances inside will release energy as heat. The air's heat-induced expansion and any solvents present generate ultrasonic waves that are measured and cataloged by the photoacoustic transducer. A neural network analyzes the signals and learns how a biological particle's "signature" is different from the sound-wave signature of inanimate airborne particles.

The first phase of testing determined wavelengths and pulse energies that were suitable, from mild spectroscopic scans to high-energy lasers capable of ablation. To isolate bioparticles and catalog their reactions to laser pulses, optical tweezers immobilized them while different photoacoustic cells and transducer geometries were tried out. So far, these tests determined that only a little water needs to be mixed with the samples to ensure adequate signal generation for the transducers. The cataloging of the various resulting "signatures" has just begun.

Sound measure

"There are sound waves given off in the ultrasound frequency range, and we use our special transducer for measuring these sound waves," said Small. "There is diversity among particles, not just in terms of very simple light-to-heat conversions, which is basically what goes on: Light is absorbed by the particles and released as heat, which causes a rapid expansion, which generates pressure waves. But also, some particles generate bubbles around them, which creates huge acoustic waves which are different for bacteria when compared to soot or dust."

The laser pulses ignite changes in the immediate electron shell of an absorbing molecule, causing the environment to "rush in" and creating vibronic energy in mere picoseconds. This acoustic energy heats the surrounding water, causing expansion and sometimes momentary fluorescence. Depending on the energy of the laser, it can also cause bond cleavages, electron ejection, proton transfer and a process called electrostriction in which solvents contract around newly formed charges, resulting in conformational changes in volume.

These separate reactions are integrated into a single ultrasonic pressure wave propagating away from the illuminated particles. Because the components cannot be easily separated, a neural network learns the overall "signature" of each sample from the volume changes sensed by a piezoelectric transducer.

The sample being tested is held in a cuvette that can be pulsed three to 20 times per second. A beam splitter enables the wavelength and relative pulse energies to be measured as they are delivered to the sample through a neutral-density filter, which can trim pulses to precise energy levels. A variable rectangular aperture feeding a cylindrical lens allows a vertical sheet of light to be focused on the sample. The ultrasonic transducer, clamped on the side of the cuvette, feeds the resulting acoustic waveforms into a digitizing oscilloscope for storage. A computer then supplies a trigger to the oscilloscope to transfer the data over a SCSI bus for analysis by neural-learning algorithms.

Noisy prototype

The 1-cm path-length cuvette is maintained at a precise temperature of 4°Centigrade (0.02°) with a layer of mineral oil between it and the piezoelectric transducer to minimize abrupt changes in acoustic impedance. The transducer is maintained at a known tension against the cuvette, so it can be removed and replaced with exactly the same pressure applied. Magnetic "stirring" flushes photo products from the illuminated area just before and after each laser pulse.

The prototype, bulky and noisy, occupies several cubic feet and sounds like a vacuum cleaner. Next-generation devices, Small said, will have more integration, but the first models will only be suitable in environments that are already loud. "Airports and other public areas will be a good place for it, because our approach is proactive — it's designed to screen the air on a long-term basis," she said.

"We are starting to work with the U.S. Postal Service to find out what specifically happens in a postal sorting facility that could lead to anthrax in the air, and we are trying to custom-tailor a device specifically for that application," said Small.

For the future, Small will next attempt to build a smaller, more-portable unit that can be carried to different locations and set up for immediate screening. She will also be working on end-user software that will allow even untrained personnel to operate the device with a goal of declaring an area "clean" in under 30 minutes.