Near to far: chemical sensors take on new importance

Whether it's cancer in a laboratory well or deadly sarin on the battlefield, detecting chemicals or harmful organisms is not a task that optical engineers take lightly. The process is complex; toxins, DNA, and tissue samples are diagnosed using labor-intensive and costly processes that require chemical tags, stains, and trained personnel. The situation is exascerbated in remote applications.

Figure 1. WILDCAT sensor production concept for JSWILD.

In response, companies and government agencies are looking at optical methods to increase the efficiency of hospital laboratories, as well as mobile military personnel hoping to avoid clouds of chemical death.

Typically, malignant or benign cancers are evaluated using stains and microscopes. The samples are first excised from the patient, sent to a central laboratory, and tested; eventually, the results are passed on to the anxious patient. For years, medical researchers have looked for better ways to evaluate the human condition, considering everything from implantable monitors with sophisticated transmission schemes to automated systems that use optical fibers with variations of absorption spectroscopy, changes in diffractive index, and other optical phenomena.

Today, companies such as Luna Innovations (Blacksburg, VA) are working on semi-invasive procedures that involve inserting a light pipe fiber with long-period gratings directly into the human body. Luna's approach uses a special process to write the long-period gratings into the fiber and then coats the fiber with receptive material geared toward the particular chemical or pathogen. As the cancer cells or chemical binds to the outside of the fiber, it interferes with the evanescent wave inside the cladding, modulating the spectral profile of propagating laser light. Similar methods, such as surface plasmon resonance, require special metal coatings, or must be thrown away after a single use.

According to Mark Jones, director of biological and chemical instrumentation at Luna Innovations, multiplexed versions of the systems sporting eight fibers in a single array and using a single source are already in beta test. "The system has been demonstrated for a variety of targets," Jones said. "We've tried it on big proteins and small molecules, such as atropine and pesticide-type targets. We're in the process of demonstrating it for DNA. We can directly detect picograms of material. Because we've got fiber optics, it has field-portable capabilities. Soon, we'll start in the field to demonstrate environmental applications."

Fiber optic probes are just one category of optical devices used in field measurements. Unlike fiber-delivered point-of-contact measurements, many field applications require a wide coverage area, such as the perimeter of a refinery or military installation, or even perform the rugged job of reconnaissance when chemical and biological threats are expected. According to several experts in the defense industry, the U.S./Iraqi war has prompted a resurgence in remote chemical and biological threat detection. As a result, several U.S. agencies continue to make improvements to commercial systems -- such as the hyperspectral imager -- as well as develop emerging systems, such as light detection and ranging (LIDAR) and differential absorption LIDAR (DIAL) and its counterpart differential scattering (DISC) in particular.

Under a program called 'Caliope,' the U.S. Department of Energy (DoE) evaluated all these systems for their effectiveness in detecting harmful chemical agents. According to Dan Beatty, project manager at the DoE's Office of Nonproliferation Research and Engineering, the good news is that all these systems work; the bad news is that there's no one system that can do it all. "Each system has its little niche," Beatty said. "That's something we knew going in, but it was reinforced. There's no silver bullet -- no one technique that can do everything we want to do -- and that complicates implementation. It means that whatever you do is going to be a compromise of technology, the wavelength region of interest, how well it works, and how much it costs."

Perhaps the most mature technology is hyperspectral imaging. A number of companies, including satellite manufacturers, are having commercial success generating huge data sets called hyperspectral cubes. Essentially, a hyperspectral image cube is the same picture viewed from different wavelengths. As the number of individual bands increases, the cube grows in size and information. Although not all hyperspectral systems are imagers (in fact, Beatty said the most sensitive systems tend not to deal in image space), both systems share at least one common attribute: gleaning the useful information from the cubic set is not easy.

"There are trade-offs between imaging and chemical sensitivity and how you interpret the data," Beatty said. "It's difficult to build a nice instrument, and more difficult to properly exploit the data. That's really our focus right now...to find better methods to exploit the hyper cubes, both quicker methods and more automated. There's probably more information in the hyper cubes than we've been able to extract so far. We're developing new algorithms to get there."

A final solution may be a combination of passive hyperspectral systems and active systems such as LIDAR and DIAL. "In raw chemical sensitivity, it's LIDAR and DIAL, but a LIDAR only looks at a very narrow point. You need to know what you're looking for to begin with. That's why hyperspectral is important -- because you can see [chemical] features in a context," Beatty said.

Steven Alejandro of the Air Force Research Laboratory said DIAL and LIDAR represent the state of the art in remote sensing of chemical and biological agents. Alejandro is working closely with his Army counterpart, Cindy Swim, of the Edgewood Research, Development, and Engineering Center, and they have divided between them the chemical systems for air and ground platforms to test a variety of new laser sources, detectors, and platform designs. "With DIAL and LIDAR, the areas of interest are basically in the mid- to long-wave (IR), and the challenges are getting a system sensitive enough while suitably packaging it for the platforms of interest," Alejandro said.

Essentially, LIDAR sends a fixed wavelength laser pulse out into a region of the atmosphere where chemical or biological agents are thought to be present. DIAL uses frequency-agile laser sources that send out a number of narrow-band pulses in a given spectral range. By performing the same measurements as the LIDAR systems, DIAL can provide the operator with a much clearer picture of the chemical's spectral fingerprint. The reflected signal is analyzed, and, based on the spectral modulation of the beam, operators can deduce what portions of the spectrum were differentially absorbed or scattered by airborne elements, in what quantity, and, therefore, the type of chemical.

Based on the differences between the two systems, LIDAR is generally considered best for long-range atmospheric mapping, such as the movement of a chemical cloud and analyzing the size and density of airborne particles, including biologic threats. DIAL is seen as the surgeon among the doctors, revealing detailed information on chemical or biologic constituents, although usually at the cost of range. Frequency-agile lasers are less efficient because of the optical parametric oscillators (OPOs) or other tuning systems that enable multiple-wavelength emission.

Another challenge to remote laser sensing is posed by water in the atmosphere. Chemicals generally absorb more light in the UV and IR than in the visible. Unfortunately, water also absorbs light and is always present in the atmosphere in greater or lesser degrees. Water does not present a problem in two specific spectral ranges, however. "We're restricted to those window regions of the atmosphere at 2 to 5 µm and 8 to 12 µm where water isn't a problem," Alejandro said. That means the source could be a Nd:YAG or YLF that uses an OPO to get to the longer wavelengths. Some are single-stage OPOs, others are dual; some are angle tuned, others are not. There are many different ways to skin that cat, and each has advantages and disadvantages."

Figure 2. FAL sensor. The WILDCAT design is based on the database obtained with the FAL sensor.

Swim's group has developed a powerful, portable sealed-CO₂-laser chemical detector, the Frequency-Agile Laser Sensor (FAL Sensor; Figure 2). The FAL was initially designed by Raytheon as part of a mobile chemical detection system with a specified range of 10 km. This first FAL, however, meets only half of the current 20-km range requirement, operating between 100 and 200 mJ/pulse in a 20-wavelength burst with 200-Hz pulse-to-pulse repetition rates at a 40 percent duty cycle (80 Hz, average). The next-generation CO₂ LIDAR, known as the Warning and Identification LIDAR Detector for Countering Agent Threats (WILDCAT), has been designed by Raytheon and STI Optronics to expand the system's range and versatility. According to Swim, the STI Optronics WILDCAT laser is a "rapidly tunable, 1 J/pulse, 100-Hz transverse-electric-discharge CO₂ laser [that] is line tunable between pulses at the 100-Hz pulse repetition frequency by using a diffraction grating and either a rotating octagon or a quasi-resonant galvanometer-driven mirror."

The new Joint Services Warning and Identification LIDAR Detector (JSWILD) program will use the WILDCAT concept (Figure 1) for expanding the required range to 20 km to protect fixed installations.

Figure 3. SHREWD concept.

"We really need 1J/pulse to get the ranges we need through the atmosphere," Swim said. "[JSWILD] is primarily for chemical detection -- the aerosols and rains in addition to vapor detection. It also is capable of predicting the surface contamination after the attack. We demonstrated that by scanning the LIDAR just above the ground to see where the rains were falling. And so all this is happening in a few seconds. We've designed WILDCAT for 99.6 percent probability of detection with no false alarms over mission duration. Also, there's the possibility of getting bio discrimination as well." In addition, a brassboard demonstrator known as the Standoff Handheld Realtime Early Warning Detector (SHREWD; Figure 3) is under development, and is based on solid state laser technology for applications that require a smaller (30-lb) system and up to 5-km range, such as individual soldier systems, unmanned aerial vehicles, helicopters, or vehicles.