Using laboratory tools for on-site analysis isn't a new idea. But recent advances in instrumentation and methods for on-site sampling and characterization are enabling scientists to collect samples and study them in ways that haven't been possible previously. Some of the latest efforts in this area of research were discussed earlier this month at Pittcon.

"Analyzing samples in situ offers a number of advantages over collecting samples and bringing them back to the lab," argued R. Timothy Short, a staff scientist at the Center for Ocean Technology at the University of South Florida (USF), St. Petersburg. For example, on-site analysis lowers the risk of sample loss, contamination, and degradation, Short explained. In addition, analytical results can be obtained more quickly--even in real time--and continuous sampling may be possible. That kind of information can be used to determine where to sample next or how to respond to hazardous conditions.

In some situations, scientists can capitalize on the benefits of on-site analysis through straightforward means--by setting up mobile field labs, for example.

But for Short and his colleagues who work in marine sciences, oceanography, and related areas, it means taking a plunge--literally. The USF scientist and his coworkers are developing methods for operating mass spectrometers underwater.

Mass spectrometry is attractive for underwater analysis because of its ability to detect trace elements, determine isotope ratios, and measure a variety of dissolved gases or complex biomolecules, Short noted. The instrument can also be used to quantify pollutants and volatile organic compounds, a capability that may be invaluable for tracking chemical spills or studying their effects, he added.

The catch is that there isn't a single mass spectrometer-sample interface configuration that's suitable for all of these types of analyses," Short acknowledged. Currently, the USF team, which also includes staff scientists Strawn K. Toler, Friso H. W. van Amerom, and others, introduces samples to the instrument through a membrane.

The system consists of an inlet device that pumps the lake water or seawater, heats the sample to a fixed temperature, and delivers the water to a membrane (polydimethylsiloxane in some instruments). Dissolved gases, organic compounds, and other analytes diffuse through the membrane, which serves as a barrier between the water and the mass spectrometer, which is held under vacuum.

Short pointed out that for applications in which the instrument is lowered hundreds of meters below the water's surface, the pressure differential across the membrane is enormous. Pressure increases approximately 1 atm for every 10 meters of depth, he said.

After passing through the membrane, analyte molecules are ionized using a conventional electron-impact ionizer and then separated and detected in a commercial 200-amu quadrupole mass filter. The entire assembly, which includes vacuum pumps, a computer, electronics, and other components, is held in a watertight, pressure-proof housing. Data are stored on the instrument's computer and can be analyzed after the system is retrieved. The signal can also be monitored remotely in real time via a communication cable or using wireless technology.

Some of the first field tests were conducted a couple of years ago in the Gulf of Mexico near Ft. Myers, Fla. In that study, a diver guided an untethered instrument to a depth of up to 20 meters to measure dissolved gases in the vicinity and interior of underwater hydrothermal vents. Among other findings, Short and coworkers observed that the carbon dioxide and oxygen signals, which varied with time and position in a complex way, mirrored one another almost exactly--meaning that the concentrations of those species were found to be related inversely. That type of correlation typically signifies biological activity, Short noted.

TAKE THE PLUNGE Want to know what's down there (besides water)? University of South Florida scientists find out by using submersible mass spectrometers, as seen here being positioned by a diver in the Gulf of Mexico (left) and being lowered into Lake Yellowstone on a remotely operated vehicle.

COURTESY OF USF CENTER FOR OCEAN TECHNOLOGY

The USF team conducted similar tests in Lake Yellowstone, Wyo. Working with colleagues from the University of Wisconsin, Milwaukee, and Eastern Oceanics, West Redding, Conn., the group deployed a tethered system in which a mass spectrometer situated at the bottom of a remotely operated vehicle was used to examine vent gases such as methane, nitrogen, carbon dioxide, and argon from several vents at depths of up to 72 meters. The researchers found distinct differences in the gas compositions at each of the sites, including evidence of hydrogen sulfide in one of the vents.

Recently, in Saanich Inlet, near Victoria, British Columbia, USF graduate students Peter G. Wenner and Ryan J. Bell sent their submersible spectrometer some 200 meters below the surface and observed wide variations in the depth profiles of several dissolved gases. Short noted that an underwater ledge in the inlet restricts water circulation, leading to anoxic conditions and the presence of oxygen-free molecules, such as hydrogen sulfide and methane.

In addition to developing new sampling methods to broaden the range of analytes that can be probed with underwater mass spectrometry, the USF team is also working on automated procedures for preparing chemical-composition contour maps. The idea is to use unmanned vehicles to deploy mass spectrometers and guide them through the water and then combine the analytical signals with global-positioning-system data. That effort, which is already under way, moves the USF scientists toward their ultimate goal of developing self-directed robotic sensors.

	DEADLY Researchers are developing fast and sensitive methods to detect sulfur- and phosphorus-based chemical warfare agents.

One focus of Amirav's research is rapid detection of highly toxic compounds. "Israel faces a severe chemical-warfare-agent threat from neighbors and terrorist organizations," Amirav stated. Yet much of the technology used at the present time to alert civilians and military personnel to the presence of poisonous substances is inadequate in several ways, he added.

For example, detectors based on ion-mobility spectrometry and other methods often generate false alarms and are unable to identify harmful agents with molecular specificity, Amirav contended. That kind of information is required to develop effective medical-response and decontamination procedures and to know when it is safe for emergency workers and others to remove protective gear such as face masks. In addition, VX, which is up to 20 times more toxic than sarin, soman, and other nerve agents, cannot be detected at medically relevant levels by today's common field detectors, he stressed.

Short	Pawliszyn	Amirav
PHOTOS BY MITCH JACOBY

TO ADDRESS these problems and related ones, Amirav's research group developed a fast, handheld GC system equipped with a pulsed-flame photometric detector (PFPD). The system operates by collecting analyte molecules in air samples on a miniature trap and then desorbing the species and delivering them to a short capillary column for GC separation.

According to Amirav, the principle of detection is similar to conventional flame photometric detection with an added dimension of time. Specifically, analyte molecules are combusted in a pulsed hydrogen flame, and the emitted element-specific light (green for phosphorus and violet for sulfur) is measured in a photomultiplier tube. The process is repeated at a frequency of a few hertz.

"The most important feature of the detector is the added dimension of time," Amirav said. He explained that hydrocarbons--because of their large negative enthalpies of combustion--undergo combustion and light emission on a millisecond timescale. In contrast, compounds containing phosphorus, including nerve agents, emit light on a longer timescale than hydrocarbons. And sulfur compounds, such as blister agents, emit light on an even longer timescale.

So by using gated signal-processing methods, in which data are recorded in separate windows of time (or channels) of just a few milliseconds each, analytes can be distinguished by their heteroatoms and separated according to their elution times. Amirav stressed that unlike other techniques, such as ion-mobility spectrometry coupled to GC, the PFPD-GC system provides two unrelated detection principles for identifying analytes selectively.

Combining the PFPD, which is an element-selective detector, with volatility-based GC separation "is an effective way to achieve unambiguous chemical identification in the field while almost totally eliminating false-positive and false-negative results," Amirav noted.

Demonstrating the method's capability, Amirav presented data from test runs showing complete separation of mixtures of chemical warfare agent simulants--compounds with properties and structures that mimic the real agents. The mixture of phosphorus and sulfur compounds was separated in 30-second runs, during which the analytes were trapped and desorbed and then separated on a GC column, and the system was cooled in preparation for the next run. Similar results were shown for a test in which an organophosphorus compound dissolved in diesel fuel was readily distinguished from the large number of potentially interfering organosulfur and other compounds in the headspace above the liquid.

Amirav reported that the system, which can be used in a continuous "sniff" mode and other modes of operation, has detection limits on the order of 3 ng per m³ for organophosphorus compounds and roughly 0.2 mg per m³ for organosulfur compounds.

WHILE SOME analytical chemists concentrate on developing new instruments or techniques for analyzing samples, Janusz Pawliszyn focuses on improving methods for collecting and preparing samples. The chemistry professor at the University of Waterloo, in Ontario, reported on his research group's efforts at integrating sample handling and analysis in field-portable analytical systems.

One approach to sample preparation for on-site investigations is to use a membrane to separate analyte molecules from the medium in which they are found--for example, air--and then trap the molecules for subsequent analysis. In the setup developed by Pawliszyn's group, the species diffuse through hollow-fiber membranes or other types of barriers and then are transported by a carrier gas to a sorbent-filled microtrap, where they are accumulated. Trapping the molecules in that way leads to increased sensitivity, which is especially useful when working with dilute samples. After a predetermined period (typically a few minutes), the concentrated sample of analyte molecules is desorbed from the trap by heating the sorbent with an electric pulse and is delivered to a GC or other instrument for analysis.

By coupling the sample-collection and concentration components to a commercial, portable GC instrument, Pawliszyn and coworkers have come up with a simple and automated system that requires no solvents for sample extraction and can be used for continuous on-site air monitoring and other applications. As an example, Pawliszyn showed data from a study in which the system was used to measure eucalyptol, a-pinene, and other emissions from eucalyptus trees over the course of a day. One drawback of the design is that increasing the sample collection and concentration time decreases the frequency at which measurements can be made.

The main thrust of the Waterloo group's research is developing small probes that are well-suited to sample collection and preparation and are convenient for delivering samples to portable analytical instruments such as gas chromatographs.

One such probe that was commercialized recently is based on a tiny fused-silica fiber that's coated with a polymeric material and housed in a GC syringe-type holder or a penlike device. The tool is used for a microextraction technique in which the coated fiber is exposed to the system under investigation, and the coating traps and concentrates volatile analytes directly without additional preparation steps. The fiber containing the trapped analytes is typically inserted into the inlet of a portable instrument such as a GC coupled to a mass spectrometer, where a carrier gas strips the analytes from the fiber coating. The species are then separated and quantified in the usual manner.

Emphasizing the flexibility of the microextraction device, Pawliszyn pointed out that the polymeric coating on the fiber can be selected according to the material's affinity for specific analytes. And not only can it be used repeatedly without special cleaning procedures, but the entire sampling and analysis process can be done quickly, he said.

Underscoring the point, Pawliszyn showed results of a residential air-quality survey in which an entire house was tested in a single afternoon using two-minute sample collections and three-minute on-site GC runs. The study revealed that the air in some of the rooms of the house was fouled by high concentrations of toluene and other gasoline components. In contrast to the microextraction technique, standard air-analysis methods involving charcoal adsorption tubes and extractions using carbon disulfide would have required several days, he said.

THE LATEST EFFORT from the Waterloo team involves a specially designed needle that features a roughly 5-mm-long section near the tip that is packed with an immobilized sorbent bed. Similar to some of the other tools developed by the chemists, the miniature needle trap is used to collect and concentrate analytes and introduce them directly into a GC or another instrument without the need for additional sample-handling steps.

To optimize the low-cost sampling device, Pawliszyn's group compared several types of packing materials, including quartz wool, molecular sieves, and particles of polydimethylsiloxane and divinylbenzene. The team also evaluated strategies for immobilizing the sorbents and experimented with single-layer and multilayer configurations. Thus far, needle traps have been used to analyze airborne particulates, aerosols, insect repellants, and diesel exhaust emissions, as well as in other applications.

Pawliszyn reported that in validation studies that compared the new method with standard published procedures for monitoring indoor air quality, the needle-based technique was found to be accurate and reproducible. He added that for GC analysis using a flame-ionization detector and 25-mL gas samples, the needle-trap method yielded detection limits of roughly 0.2 ng per L for benzene and approximately 1 ng per L for o-xylene.

From underwater mass spectrometers that can probe the depths of the Pacific Ocean, to above-water headspace samplers that collect analytes on miniature traps, chemists continue to make progress in on-site sampling and analysis. Whether it's everyday samples from familiar locations or exotic species at faraway sites, the nagging concerns regarding sample quality and timeliness of results are mitigated by taking the analytical show on the road.