Capillary-attached gas sensor system design for electronic nose application

• Main Author: Bahraminejad, Behzad
• Format: Thesis
• Language: English
• Published: 2011
• Subjects: Detectors; Gas - Odorizing; Gas industry - Safety measures
• Online Access: http://psasir.upm.edu.my/id/eprint/33950/1/FK%202011%2032.pdf; http://psasir.upm.edu.my/id/eprint/33950/
• Institution: Universiti Putra Malaysia

Since the 1980s, researchers have spent a great deal of effort developing the electronic nose (e-nose) as an artificial olfaction instrument to mimic the mammalian sense of smell. Currently, the e-nose is employed in a wide range of applications, such as food and beverage quality control, environmental monitoring and medical diagnosis. Although published results have indicated great achievement in the field of gas identification systems, fabricating a comprehensive instrument for rapid and real-time gas identification is still difficult and complicated. Most of the efforts to improve electronic nose performance have been in applying diverse gas sensor technologies and signal processing techniques to increase the selectivity of the e-nose. However, the classic structure of the e-nose has remained relatively untouched. Moreover, all the detection parameters are extracted based on the instant trace of the target gas on the sensing element. In this dissertation, the design and development of a gas identification system based on a developed sensor structure and transient response analysis was studied. An air-filled capillary was attached to a metal oxide semiconductor gas sensor to fabricate a developed structure and to increase the identifying information of the gas sensor output. The new structure involved a unique gas diffusion coefficient for each target gas as an additional parameter in the transient response of the gas sensor in the reaction of a target gas. A prototype sensor based on the proposed structure was fabricated, and its physical characteristics were investigated. The dimensional effects of new physical attachment, in both length and diameter, were studied, and optimized dimensions were selected. Application of the fabricated prototype gas sensor in detecting five hydrocarbon gases was studied. The results indicated that the sensor successfully distinguished between different applied gases. The ability of the optimum sensor structure in real-time gas identification was also investigated. It proved that the early portion of the transient response could be applied to achieve a high accuracy classification performance. Then, the new sensor structure was applied to the design and development of an e-nose. The advantages of the designed e-nose were studied by applying diverse simple and complex odors. The results indicated that the new electronic nose could detect different types of odors with a high classification rate. Therefore, the development of the identifying abilities of the e-nose is another advantage of the new sensor structure. For both the single sensor and e-nose system, different compatible processing and analyzing methods were employed regarding the sensor transient response characteristics. Different pattern recognition and classification methods were also studied to generate the optimized identification unit. The new structure also reduced the interfering parameters, a common problem reported in research, and increased the reproducibility of the system. The ability of the developed sensor structure in single and mixed gas identification can be applied to fabricate smarter gas detection instruments and to employ the e-nose in gas chromatography or mass spectrometry as challenging applications for future work.