Chemosense
BACK ISSUES OF OUR JOURNAL ARE AVAILABLE FOR DOWNLOAD
Continuous real-time monitoring heralds a generational shift in pollution control
Graham Bell
CEO, E-Nose Pty Ltd
g.bell@atp,com.au
Why do some industries make such obnoxious odours? What is this costing them and us? Do they know what they are doing? What can be done about it? These are four of many questions on the issue of air quality in practically all city and country towns in Australia and elsewhere.
What many industries do, is of necessity, a smelly or chemically noxious business. Their livelihoods depend on the use of volatile chemicals or they produce them as by-products or waste. They often establish their plants and facilities a long way out of town, but, with time and “progress”, the suburbs eventually reach them, and that is usually when complaints begin.
Individual complainants, however, are no longer the only cause for concern by the industries. Growing community awareness, of the wider effects of pollution and resource usage, are bring pressures to bear on an increasing number of industries, including those that are remote from urban communities.
Complaints very often concern smells from organic activities, such as farming and food processing. Among the most “celebrated' are the smells from animal product processing, and from raising of animals in confined spaces: such as cattle feedlots, piggeries and chicken sheds. Among the top stink-makers are sewerage and waste treatment plants. Many other industries release chemicals that are not quite as unpleasant, but may be of far greater concern. Examples of those, which release solvents into the atmosphere: from the small vehicle spray painter to large plants, such as metal smelters or oil refineries. Closer to home there is the motor car or small business truck contributing to the reduced quality and safety of the air we breathe.
The costs include the individual's health and discomfort, and the time taken to make and process the complaint. Affected industries acknowledge that complaints cost millions every year. They have to devote valuable people and resources to investigating the issues raised, documenting and recording their responses, dealing with possible and real litigation, and complying with orders from authorities such as an environmental protection agency (EPA). These events can result in costly action and unplanned expenditures.
At the national level the cost adds up to billions of dollars. Part of this is the price an industrialised society must now pay for operating in a clean and healthy environment and sustaining its long-term future. The rest, probably the greater proportion, is arguably a waste of resources that produces nothing. It is this fraction that can be reduced by better use of technology, and the savings redeployed to other ends.
Industrialists know they have a problem, but only in general. Specifically, each situation is made complex by the intermittent nature of emissions; variations in their quality and intensity; and the vagaries of external factors such as temperature, humidity, wind speed and direction. People at the plant adapt to the smell, and for them personally, the problem simply doesn't exist. By the time a smell is investigated the situation has usually changed: usually it has abated or changed in quality.
The current methods of establishing the strength, quality and source of the environmental odour are very difficult to carry out rapidly: they consist of drawing air into clean non-adsorbent plastic bags, from various parts of the site or downwind from it. The air is then taken, within 30 h, to a lab where chemical analysis is performed or a panel of humans assess it.
The trained human panel of 6 to 8 people sniff systematically diluted samples of the original air sample, to determine how many times a unit volume of that air has to be diluted before it can no longer be detected. These numbers are called “odour units” (OUs). Each bag sampled at any one time produces a single odour unit number. The higher the number the more strongly smelling the air is assumed to be. The effort and cost required to obtain these numbers means that not many numbers can be obtained, either over time or over a wide area, and that rapid response to a complaint or timely action to minimise a possible complaint is practically impossible.
What is needed is a continuous monitoring technology with sufficient sensitivity, reliability (reproducible results) and validity (measuring what it claims to measure). Srivastava and Levy (2002) defined the optimal characteristics for an air quality monitoring system within constraints imposed by each situation's factors, such as likely concentration range of pollutants; background emissions from other sources; meteorological and geographic conditions; and measurement and calibration frequency of the system. The criteria are, briefly:
1. Sensitivity. It must meet the concentration ranges required for the job and be able to discern one quality of odour from another
2. Reliability. It must consistently produce accurate, precise, specific and reproducible results
3. Temporal Resolution. It must have a sufficiently short time period over which it makes a determination
4. Robustness. It must have a low failure or fault rate and its performance must be steadfast against such destructive influences as extremes of temperature, dust, wind movement and humidity (including rain or snow).
In addition, a monitoring system should be affordable, of convenient size, draw minimal power and communicate efficiently with its operator.
Inspired by biological systems, a technology is emerging from the pure science of the olfactory system (see review of possibilities in Bell, 1996). Novel surfaces which capture different species of molecule have been combined into arrays that mimic the biological nose to produce a unique, albeit fuzzy, representation of a complex mixture of volatiles. Various chemical sensor arrays (electronic or e-noses) have been successful, to varying degrees, at meeting the above criteria for detection and monitoring of airborne chemicals.
A growing number of labs and companies have become involved in recent years. A cursory web search on www.google.com for the words “electronic nose” will yield at least 143,000 references. There are currently between 20 and 30 labs and companies working on e-noses. Interest in technical and intellectually challenging problems associated with e-noses is manifested in the lively web user groups and a growing number of international conferences.
Companies producing e-noses for commercial applications include Cyrano Sciences (Pasadena, CA) who have targeted medical diagnostic applications and screening of food and packaging quality, with a hand-held device called “Cyranose”. Marconi Technologies (Chelmsford, Essex, UK) has used e-noses to detect gas leaks, to monitor the quality of propylene glycol in lotions, and freshness in frozen shrimp. AromaScan, (Crewe, Cheshire, UK) has placed over 200 electronic noses into laboratories around the world, including one reportedly installed on the Mir space station to detect odours from failing electronic components (Skelley, 2000).
A new generation of electronic chemical monitoring devices (e-noses) promises to change this situation for the good of all concerned. The arrays are usually smaller (fewer sensors) than that shown in the illustration and target specific problem odours. The kind of odour, its context and its chemical composition are studied first, and then an array is specified, that best responds to the priority odour. The specialised e-nose responds at levels that give rise to complaints, against and discernible from characteristic background odours.
These e-noses can be readily combined as multi-unit networks to provide wide area or very long perimeter monitoring of a large installation or zone of interest. Outdoor security and military applications are areas of potential application in which these new generation sensors are likely to appear in the near future.
The development of the internet, geo-positioning and communications by satellite makes the retrieval of information possible from multiple e-noses deployed extremely remotely. The economics of scale possible in the electronics industry dictates that cost-per-unit will become trivial and thereby allow networks of limited-life, self-disposable e-noses to be deployed by the thousand. Operating when visual and vibrational detectors may be ineffective, these new electronic noses may play a vital part in surveillance along a country's border; in difficult-to-reach mountainous areas; or even across expanses of ocean.
The amount of odour data that can be gathered over an expanse of terrain makes mapping of flow dynamics from industrial sites into areas of complaint or concern possible at a level of accuracy only ever dreamed of by pollution modellers.
The Australian company, E-Nose Pty Ltd, a spin-off of UNSW (see Hibbert and Barnett, 2002), now has a range of relatively small, specialised e-nose “sentinels” that are tailored for specific odour environments such as abattoirs and sewerage treatment plants and pumping stations. The company has also developed a novel self-diagnostic and calibration system to add-on, when appropriate, to each e-nose. The combinations of sensors is selected to do the job needed by the client, with resulting robust and reliable operations at levels of sensitivity well within the range of reported nuisance odours.
The systems were trialed in industrial settings, in “hostile” (temperature, humidity and particulates) indoor and outdoor settings, and proved to be very robust, meeting adequately all four criteria listed by Srivastava and Levy (2002).
The system was calibrated against human “odour unit” measurements and provided useful data down to 1 OU. The speed at which this human-calibrated data can be gathered and analysed heralds a generational shift in odour measurement and monitoring.
It makes possible realistic mapping of the dispersion of odours from industrial sites at different times and in different settings, thereby bringing clarity to these issues. This will be a “win-win” for the industrialists and the community.
Remote access by internet allowed the operator (and the central development lab) to observe the dynamic status of the odours emitted by his plant at a rate of more than one measurement per second. They could observe at any time, the current and past performance of the plant over the past days and weeks.
In addition, the company developed software to predict the likelihood that the plant will exceed community complaint levels in the coming 30 minutes. This was achieved with a reliability coefficient of 0.96.
The savings in costs to the industries, once the technology is adopted, will be massive. They will be constantly equipped with information, and will manage their odour abatement investments more efficiently. Freed from constant threat of litigation from the community and EPAs, they will be able to get on and grow their businesses without relocating and will have newfound resources to apply to better things.
E-Nose Pty Ltd would like to hear from anyone interested in assisting it in commercialising its inventions by installing prototype systems into their industrial settings (contact g.bell@atp.com.au). The products are scheduled to be launched commercially early in 2005.
Bell, G.A. (1996) Molecular mechanisms of olfactory perception: Their potential for future technologies. Trends in Food Science & Technology, 7, 425 - 431.
Hibbert, D.B. and Barnett, D. (2002) Chemical Sensor and E-Nose Program. ChemoSense, 4(3), 8-10.
Srivastava, A.K. and Levy, D.C. (2002) Gas sensor monitoring of environmental air quality, ChemoSense, 4(3), 8-10.
Skelley, D.S., (2000) www.devicelink.com/ivdt/archive/00/01/004.html
