Loading...
Projects / Programmes source: ARIS

Integrated multi-channel artificial nose for vapor trace detection

Research activity

Code Science Field Subfield
2.09.03  Engineering sciences and technologies  Electronic components and technologies  Microelectronics 

Code Science Field
T171  Technological sciences  Microelectronics 

Code Science Field
2.02  Engineering and Technology  Electrical engineering, Electronic engineering, Information engineering 
Keywords
Vapour trace detection, capacitive micro-sensors, machine-learning, chemical selectivity, MEMS, COMB sensor, sensor arrays, TNT.
Evaluation (rules)
source: COBISS
Researchers (16)
no. Code Name and surname Research area Role Period No. of publicationsNo. of publications
1.  38444  Daniele Bertocchi  Engineering sciences and technologies  Researcher  2017 - 2020 
2.  33297  Mitja Brlan    Technical associate  2017 - 2020 
3.  36635  PhD Helena Brodnik  Chemistry  Researcher  2017 - 2020  60 
4.  29523  PhD Anton Gradišek  Physics  Researcher  2017 - 2020  441 
5.  24914  Carmen Hladnik Prosenc    Technical associate  2017 - 2020 
6.  03321  Ivan Kvasić  Physics  Technical associate  2017 - 2020  22 
7.  03494  PhD Marijan Maček  Electronic components and technologies  Retired researcher  2017 - 2020  178 
8.  03426  MSc Bojan Marin  Physics  Technical associate  2018 - 2020  93 
9.  09089  PhD Igor Muševič  Physics  Head  2017 - 2020  751 
10.  24326  PhD Aleksander Sešek  Electronic components and technologies  Researcher  2017 - 2020  142 
11.  00166  PhD Drago Strle  Electronic components and technologies  Researcher  2017 - 2020  244 
12.  12338  PhD Miha Škarabot  Physics  Researcher  2017 - 2020  253 
13.  18423  PhD Bogdan Štefane  Chemistry  Researcher  2017 - 2020  401 
14.  10476  PhD Janez Trontelj ml.  Electronic components and technologies  Researcher  2017 - 2020  60 
15.  11035  PhD Aleksander Zidanšek  Physics  Researcher  2019 - 2020  361 
16.  28235  PhD Erik Zupanič  Physics  Researcher  2017 - 2020  131 
Organisations (3)
no. Code Research organisation City Registration number No. of publicationsNo. of publications
1.  0103  University of Ljubljana, Faculty of Chemistry and Chemical Technology  Ljubljana  1626990  23,192 
2.  0106  Jožef Stefan Institute  Ljubljana  5051606000  91,025 
3.  1538  University of Ljubljana, Faculty of Electrical Engineering  Ljubljana  1626965  27,847 
Abstract
This project is targeting an entirely new line of scientific research and development by applying new methods of artificial intelligence to a newly developed 64 channels microcapacitive sensor array for detection of vapour traces of hazardous molecules in the atmosphere. There is a tremendous activity in developing new methods and concepts of artificial nose for obvious reasons of preventing threats to the society. During the past ten years, many different sensors and concepts were developed for the detection of very low atmospheric levels of explosives, such as TNT, PETN and RDX, with the aim of securing urban environment against terrorist threats. Presently, the level of sensitivity of state-of-the-art sensor systems is sufficient to detect a very low concentration of targeted molecules of one molecule in 1012 to 1014 molecules of the atmosphere. We have demonstrated through the successful completion of several research projects (Strle et al., IEEE Sens. J., 2012) that microcapacitive sensor arrays based on COMB microcapacities and low noise detection electronics are able to detect such low levels of dangerous vapours. Whereas the sensitivity we have achieved is at the cutting edge of today’s sensor technology, the shortcomings are in poor chemical selectivity of sensor arrays formed of up to 16 chemically differently sensitive microcapacitors. The objective of the present proposal is to enhance the methods of signal processing and pattern recognition using a newly designed and developed 64-channel sensor array. This will be a small but important step forward towards the realisation of future high-density electronic noses, which are expected to contain thousands of sensing pixels in the near future, thus mimicking the chemical selectivity of a dog’s nose. The proposed research aims at showing proof of concept of applying methods of artificial intelligence and in particular machine-learning to enhance the chemical selectivity of a 64 sensor array. This will be done by major development of the new electronic nose based on a new sensor array consisting of 64 chemically differently functionalised sensors and by performing a huge number of sensing experiments, where a database will be filled with the response of this array to different and well defined mixtures of substances in the atmosphere. Algorithms of artificial machine-learning will be developed to make a decision on the presence of hazardous targeted molecules in the measured volume of the atmosphere. This project will be conducted for the first time by an interdisciplinary research team formed of physicists, organic chemists, microelectronic engineers, mathematicians and artificial intelligence experts. A well-defined research plan, including risk identification and mitigation, has been constructed. We trust a successful completion of this project will significantly increase the technological readiness of the artificial nose, which is also a point of interest of Slovenian SME RLS d.o.o., who strongly supports this project.
Significance for science
We propose a very ambitious research project which aims at increasing the chemical selectivity of sensing vapour traces of hazardous matter in the atmosphere. We are targeting two challenges: (i) to increase the number of chemically differently functionalised microsensors up to 64 sensors forming an array, (ii) to apply machine-learning methods of artificial intelligence to process the multitude of captured data from the sensor array. The successful completion of this research project will not only push the detection limits and sensitivity of today’s state-of-the-art sensorial systems, but will also add the missing functionality to today’s sensors, which is the chemical selectivity. This will open a completely new line of possibilities for molecular recognition by combining cutting edge microelectronic design and solutions in sensor arrays with methods of artificial intelligence. To our knowledge, no attempts have been made up to date of using machine-learning to manage and improve chemical selectivity of sensors arrays, because no such sensors are available yet. High sensitivity, high dynamic range and chemical selectivity supported by artificial intelligence would open new horizons in chemical recognition platform, as it would be possible to discern specific details in sensor response of the entire array. This is a multidisciplinary project that will be conducted by a well-organised team of physicists, organic chemists, microelectronic experts, mathematicians and artificial intelligence specialists. The team has a long history of collaboration and the artificial intelligence methods will be incorporated for the first time. The reason for this is that the hardware technology has matured to the level where input from the soft sciences is needed.
Significance for the country
The proposed project integrates research efforts and ideas of three different groups from three distinct scientific disciplines: physics, chemistry and electronic engineering. Each of these disciplines has a long and fruitful tradition in Slovenia, but cooperative projects, where all three disciplines would cooperate, are extremely rare. The three research groups have in the past successfully cooperated in the classified research project UNos (Artificial Nose), financed by the Slovenian Ministry of Defence. The results of this project are at the cutting edge of today's science and technology. During the period of seven years, the research groups located at the Jozef Stefan Institute, Faculty of Mathematics and Physics and Faculty of Chemistry, both at the University of Ljubljana, developed a very tight collaboration that resulted in very good project output. From this solid background, the proposed project introduces important new aspects of molecular detection and recognition, which are presently among the most relevant in the international scientific community. Investigations in new, highly relevant topics and coverage of wide range of research fields are prerequisites for successful continuation of the research tradition in Slovenia and for the organisation of the supporting educational programmes at the University of Ljubljana. Of particular importance is the interest in our research project, demonstrated by several international companies: Infineon Technologies Austria AG, ST Microelectronics Italia and Thales Group France, which adds an important dimension of international relevance to the proposed work  that might continue in a form of a future European project, or some other form of international research cooperation. The proposed research project has no direct impact on a particular company, however we expect in time the development of smaller-scale industry or high-tech companies that could exploit and use our technological and research achievements. The results of the proposed research are the baseground for the development of different miniaturized sensors that can be used in different fields. Because of this reasons several domestic and foreign companies have expressed very strong interest. In addition, several small domestic companies and international investors have expressed the interest for further development and commercialization of the results. We should stress that this project is strongly supported by the fastest growing Slovenian SME RLS merilni instrumenti d.o.o. The letter of support of this company towards our proposal is included in this submission.  A successful completion and further commercialized of the proposed project would have a direct impact to the domestic economy because of the creation of new high-tech jobs, which would be beneficial for long term Slovenian economy growth.
Most important scientific results Interim report, final report
Most important socioeconomically and culturally relevant results Final report
Views history
Favourite