Russell Parris, chief development officer, Owlstone plc, Cambridge, UK.
Electronic noses for drug- and explosives-detection have been under development for years, and ion mobility spectrometry has been part of the toolbox for about a decade. Working for the Federal Aviation Administration, Sandia National Laboratories in 1999 developed a portable IMS unit that required a chemical preconcentrator unit. Four years later, the system had shrunk to the size of a toolbox. Compared to the Owlstone’s nano device, that’s clunky: Its sensing chip is less than 1cm long. The only other components are a second silicon chip for the electronics and an ionization source. This chemical sensor-fabricated with lithographic techniques, with components etched on a silicon substrate-is hundreds of times smaller than Sandia’s IMS unit. It also costs a fraction as much: Sandia’s sniffer and preconcentrator was estimated to cost $20,000-$40,000. Owlstone’s sensor can be fabricated for $5.
A reliable, miniature sensor that can detect chemical components in parts-per-billion concentrations has many applications, and engineering researchers at the four-year-old firm are working with systems integrators and OEMs to incorporate FAIMS in a wide variety of products. Among them is Lonestar, a system for use in food and beverage production. Helping to commercialize the system is Russell Parris, Owlstone’s chief development officer. An analytical chemist, Parris joined the firm in 2005 after a year at Fort Halstead, where he was part of a multidisciplinary team working on detection systems for chemical threats to military and government offices. He holds a doctorate degree in instrumentation and analytical science from the University of Manchester.
The chip in Owlstone’s sensor is less than 1cm long; the entire sensor easily fits in a hand’s palm. Reliability and low-cost are part of the technology’s appeal for food and beverage sensor challenges. Source: Owlstone plc.
Parris: In Ministry of Defense work, I was competing against a dog. I could never duplicate the detection capabilities of a well-trained dog, but I could beat him if he had a bad day. In our food work, we’re often trying to replace or at least reduce the number of human tasters. Most error in odor evaluations is introduced by human bias.
FE: How does FAIMS differ from the mass spectrometer using gas chromatography?
Parris: The gas chromatography mass spectrometer is the gold standard. We are not as good as that technology, but those are lab-based instruments. FAIMS is an in-line tool. Only in the last seven or eight years have scientists begun applying it to applications like food. There are three companies worldwide involved in the effort.
FE: What does the “field asymmetric” reference mean?
Parris: We use an RF waveform to create high and low electrical fields, causing the ions to zig-zag through the chamber, instead of going straight. The advantage of an asymmetric path is that an ion spends more time in the drift cell, increasing the resolution of the chemical analysis.
FE: Some food companies use biosensors for chemical detection. How does FAIMS compare?
Parris: Those electronic noses are all right for analyzing a single gas, but we can analyze a whole range of chemicals without changing the sensor. Also, biosensors are not very stable and tend to drift over time. We beat that technology hands down. FAIMS is very good at discriminating between chemical compounds, minimizing the problem of false positives, and the solid-state sensor has no moving parts, which helps it cope with the industrial environment.
FE: How close are you to delivering a commercially viable system?
Parris: Food companies are evaluating six of our instruments in the field, with 10 more on order. It’s looking very good, though we’re still seeking that killer application in flavor, aroma and taint detection. But ease of use and in-line detection are significant advantages. Standard analytical techniques require trained operators, are expensive and do not deliver real-time results.
Over the decades, users of mass spectrometry have built up a huge library of chemical fingerprints. FAIMS is a very small community, and we are still developing the library. We have to take the system into the field and force failure to be sure the technology works. We went to a Tennessee sausage factory to field test FAIMS’ ability to identify markers for biogenic amines in decomposing pork. Continuous sampling of the headspace of sample vials was done at different storage temperatures over 48 hours, using clean, dry air flushed through at a rate of 90ml a minute. The experiment produced a positive ion-mode color map of decomposition markers and a positive test of an instrument that can diagnose food freshness and detect spoilage.
FE: You have explored placing the sensor inside fermentation tanks for beer and wine. What is the advantage?
Parris: Fermentation produces the chemical diacetyl, one of the compounds used to create artificial butter flavor. Buttery flavored beer or wine is undesirable, and breweries and wineries often over-ferment their products to make sure the diacetyl is fully dissipated before bottling. There’s sufficient airflow and headspace in a fermentation tank to position a sensor and wire it directly to a PLC for chemical analysis. The analysis would allow producers to cut batch storage time and optimize production.
FE: What are the sensor’s limiting factors?
Parris: Airflow is necessary to drive ions through the cell, or else they will lose their charge and be useless. We’re looking at incorporating an ion pump to drive ions through the chip and decrease the airflow rate. The requirement for a power supply is a constraint, but this should not seriously constrain portability in food applications.
On the plus side, the sensor head is tiny, and the amount of ambient air needed to get a chemical sample is likewise small. The length of the path in the chamber is about a millimeter long, and there are 12-15 path lengths in each chip. The chip is the chamber. Because the chip is so small, we can create much higher electrical fields without causing a breakdown of air.
Some gases are easier than others to ionize, and a UV source works with them. For others, Nickel-63 is the source. Nickel-63 is low-grade radioactive material, and that always worries some people. But in fact, the radiation is very low, and the risk of exposure is negligible.
FE: The other limitation is that the target chemical must be known-the electronic signature needs to be in the library, correct?
Parris: Yes, you have to train the system to look for a particular chemical or compound, but this takes seconds and is very simple to execute. As long as there is a volatile marker to identify an ingredient or a residual, the system can detect it.
FE: How accessible is the technology for food-plant use?
Parris: Much of our current work is focused on limiting the technology’s functionality and helping to devise applications that don’t require any interpretation. We’re doing the hard work of developing the algorithms that analyze the sensor’s readings and generate a red light/green light result. If users have to think about the system, they’re not going to use it. They simply want a response that confirms a product is good or bad.
The complexity of the raw data can be remarkable. We’ve analyzed odor characteristics of different brands of ketchup, for example, and distinct signatures were found because the variety and concentration of the ions varies. We’re not sure what chemicals we’re picking up, but distinct signatures of regular and garlic-flavored Heinz ketchup and Tesco ketchup can be mapped. This could be useful in verifying the authenticity of a branded product.
FE: What did you apply FAIMS to in your work for the military?
Parris: Explosive devices and chemical agents are two of the biggest threats to defense forces. Explosives analysis is the toughest detection challenge because the chemicals are not very volatile when stored. Bulky equipment is needed.
Chemical weapons are another matter, and FAIMS has considerable potential. The technology could create a button-sized detector that would be worn on the lapel of every soldier on a fighting front. Tiny battery-powered, solid-state sensors could be provided to every serviceman and woman to create a network of sensors to alert central command of not only the threat of nerve gases and other toxins but also where the greatest concentrations of those chemicals can be found.
