SSeennssoorrss && TTrraannssdduucceerrss Volume 126, Issue 3, March 2011 www.sensorsportal.com ISSN 1726-5479 Editors-in-Chief: professor Sergey Y. Yurish, tel.: +34 696067716, fax: +34 93 4011989, e-mail: editor@sensorsportal.com Editors for Western Europe Meijer, Gerard C.M., Delft University of Technology, The Netherlands Ferrari, Vittorio, Universitá di Brescia, Italy Editor South America Costa-Felix, Rodrigo, Inmetro, Brazil Editor for Eastern Europe Sachenko, Anatoly, Ternopil State Economic University, Ukraine Editors for North America Datskos, Panos G., Oak Ridge National Laboratory, USA Fabien, J. Josse, Marquette University, USA Katz, Evgeny, Clarkson University, USA Editor for Asia Ohyama, Shinji, Tokyo Institute of Technology, Japan Editor for Asia-Pacific Mukhopadhyay, Subhas, Massey University, New Zealand Editorial Advisory Board Abdul Rahim, Ruzairi, Universiti Teknologi, Malaysia Ahmad, Mohd Noor, Nothern University of Engineering, Malaysia Annamalai, Karthigeyan, National Institute of Advanced Industrial Science and Technology, Japan Arcega, Francisco, University of Zaragoza, Spain Arguel, Philippe, CNRS, France Ahn, Jae-Pyoung, Korea Institute of Science and Technology, Korea Arndt, Michael, Robert Bosch GmbH, Germany Ascoli, Giorgio, George Mason University, USA Atalay, Selcuk, Inonu University, Turkey Atghiaee, Ahmad, University of Tehran, Iran Augutis, Vygantas, Kaunas University of Technology, Lithuania Avachit, Patil Lalchand, North Maharashtra University, India Ayesh, Aladdin, De Montfort University, UK Azamimi, Azian binti Abdullah, Universiti Malaysia Perlis, Malaysia Bahreyni, Behraad, University of Manitoba, Canada Baliga, Shankar, B., General Monitors Transnational, USA Baoxian, Ye, Zhengzhou University, China Barford, Lee, Agilent Laboratories, USA Barlingay, Ravindra, RF Arrays Systems, India Basu, Sukumar, Jadavpur University, India Beck, Stephen, University of Sheffield, UK Ben Bouzid, Sihem, Institut National de Recherche Scientifique, Tunisia Benachaiba, Chellali, Universitaire de Bechar, Algeria Binnie, T. David, Napier University, UK Bischoff, Gerlinde, Inst. Analytical Chemistry, Germany Bodas, Dhananjay, IMTEK, Germany Borges Carval, Nuno, Universidade de Aveiro, Portugal Bousbia-Salah, Mounir, University of Annaba, Algeria Bouvet, Marcel, CNRS – UPMC, France Brudzewski, Kazimierz, Warsaw University of Technology, Poland Cai, Chenxin, Nanjing Normal University, China Cai, Qingyun, Hunan University, China Campanella, Luigi, University La Sapienza, Italy Carvalho, Vitor, Minho University, Portugal Cecelja, Franjo, Brunel University, London, UK Cerda Belmonte, Judith, Imperial College London, UK Chakrabarty, Chandan Kumar, Universiti Tenaga Nasional, Malaysia Chakravorty, Dipankar, Association for the Cultivation of Science, India Changhai, Ru, Harbin Engineering University, China Chaudhari, Gajanan, Shri Shivaji Science College, India Chavali, Murthy, N.I. Center for Higher Education, (N.I. University), India Chen, Jiming, Zhejiang University, China Chen, Rongshun, National Tsing Hua University, Taiwan Cheng, Kuo-Sheng, National Cheng Kung University, Taiwan Chiang, Jeffrey (Cheng-Ta), Industrial Technol. Research Institute, Taiwan Chiriac, Horia, National Institute of Research and Development, Romania Chowdhuri, Arijit, University of Delhi, India Chung, Wen-Yaw, Chung Yuan Christian University, Taiwan Corres, Jesus, Universidad Publica de Navarra, Spain Cortes, Camilo A., Universidad Nacional de Colombia, Colombia Courtois, Christian, Universite de Valenciennes, France Cusano, Andrea, University of Sannio, Italy D'Amico, Arnaldo, Università di Tor Vergata, Italy De Stefano, Luca, Institute for Microelectronics and Microsystem, Italy Deshmukh, Kiran, Shri Shivaji Mahavidyalaya, Barshi, India Dickert, Franz L., Vienna University, Austria Dieguez, Angel, University of Barcelona, Spain Dimitropoulos, Panos, University of Thessaly, Greece Ding, Jianning, Jiangsu Polytechnic University, China Djordjevich, Alexandar, City University of Hong Kong, Hong Kong Donato, Nicola, University of Messina, Italy Donato, Patricio, Universidad de Mar del Plata, Argentina Dong, Feng, Tianjin University, China Drljaca, Predrag, Instersema Sensoric SA, Switzerland Dubey, Venketesh, Bournemouth University, UK Enderle, Stefan, Univ.of Ulm and KTB Mechatronics GmbH, Germany Erdem, Gursan K. Arzum, Ege University, Turkey Erkmen, Aydan M., Middle East Technical University, Turkey Estelle, Patrice, Insa Rennes, France Estrada, Horacio, University of North Carolina, USA Faiz, Adil, INSA Lyon, France Fericean, Sorin, Balluff GmbH, Germany Fernandes, Joana M., University of Porto, Portugal Francioso, Luca, CNR-IMM Institute for Microelectronics and Microsystems, Italy Francis, Laurent, University Catholique de Louvain, Belgium Fu, Weiling, South-Western Hospital, Chongqing, China Gaura, Elena, Coventry University, UK Geng, Yanfeng, China University of Petroleum, China Gole, James, Georgia Institute of Technology, USA Gong, Hao, National University of Singapore, Singapore Gonzalez de la Rosa, Juan Jose, University of Cadiz, Spain Granel, Annette, Goteborg University, Sweden Graff, Mason, The University of Texas at Arlington, USA Guan, Shan, Eastman Kodak, USA Guillet, Bruno, University of Caen, France Guo, Zhen, New Jersey Institute of Technology, USA Gupta, Narendra Kumar, Napier University, UK Hadjiloucas, Sillas, The University of Reading, UK Haider, Mohammad R., Sonoma State University, USA Hashsham, Syed, Michigan State University, USA Hasni, Abdelhafid, Bechar University, Algeria Hernandez, Alvaro, University of Alcala, Spain Hernandez, Wilmar, Universidad Politecnica de Madrid, Spain Homentcovschi, Dorel, SUNY Binghamton, USA Horstman, Tom, U.S. Automation Group, LLC, USA Hsiai, Tzung (John), University of Southern California, USA Huang, Jeng-Sheng, Chung Yuan Christian University, Taiwan Huang, Star, National Tsing Hua University, Taiwan Huang, Wei, PSG Design Center, USA Hui, David, University of New Orleans, USA Jaffrezic-Renault, Nicole, Ecole Centrale de Lyon, France Jaime Calvo-Galleg, Jaime, Universidad de Salamanca, Spain James, Daniel, Griffith University, Australia Janting, Jakob, DELTA Danish Electronics, Denmark Jiang, Liudi, University of Southampton, UK Jiang, Wei, University of Virginia, USA Jiao, Zheng, Shanghai University, China John, Joachim, IMEC, Belgium Kalach, Andrew, Voronezh Institute of Ministry of Interior, Russia Kang, Moonho, Sunmoon University, Korea South Kaniusas, Eugenijus, Vienna University of Technology, Austria Katake, Anup, Texas A&M University, USA Kausel, Wilfried, University of Music, Vienna, Austria Kavasoglu, Nese, Mugla University, Turkey Ke, Cathy, Tyndall National Institute, Ireland Khelfaoui, Rachid, Université de Bechar, Algeria Khan, Asif, Aligarh Muslim University, Aligarh, India Kim, Min Young, Kyungpook National University, Korea South Ko, Sang Choon, Electronics. and Telecom. Research Inst., Korea South Kockar, Hakan, Balikesir University, Turkey Kong, Ing, RMIT University, Australia Kotulska, Malgorzata, Wroclaw University of Technology, Poland Kratz, Henrik, Uppsala University, Sweden Kumar, Arun, University of South Florida, USA Kumar, Subodh, National Physical Laboratory, India Kung, Chih-Hsien, Chang-Jung Christian University, Taiwan Lacnjevac, Caslav, University of Belgrade, Serbia Lay-Ekuakille, Aime, University of Lecce, Italy Lee, Jang Myung, Pusan National University, Korea South Lee, Jun Su, Amkor Technology, Inc. South Korea Lei, Hua, National Starch and Chemical Company, USA Li, Genxi, Nanjing University, China Li, Hui, Shanghai Jiaotong University, China Li, Xian-Fang, Central South University, China Li, Yuefa, Wayne State University, USA Liang, Yuanchang, University of Washington, USA Liawruangrath, Saisunee, Chiang Mai University, Thailand Liew, Kim Meow, City University of Hong Kong, Hong Kong Lin, Hermann, National Kaohsiung University, Taiwan Lin, Paul, Cleveland State University, USA Linderholm, Pontus, EPFL - Microsystems Laboratory, Switzerland Liu, Aihua, University of Oklahoma, USA Liu Changgeng, Louisiana State University, USA Liu, Cheng-Hsien, National Tsing Hua University, Taiwan Liu, Songqin, Southeast University, China Lodeiro, Carlos, University of Vigo, Spain Lorenzo, Maria Encarnacio, Universidad Autonoma de Madrid, Spain Lukaszewicz, Jerzy Pawel, Nicholas Copernicus University, Poland Ma, Zhanfang, Northeast Normal University, China Majstorovic, Vidosav, University of Belgrade, Serbia Marquez, Alfredo, Centro de Investigacion en Materiales Avanzados, Mexico Matay, Ladislav, Slovak Academy of Sciences, Slovakia Mathur, Prafull, National Physical Laboratory, India Maurya, D.K., Institute of Materials Research and Engineering, Singapore Mekid, Samir, University of Manchester, UK Melnyk, Ivan, Photon Control Inc., Canada Mendes, Paulo, University of Minho, Portugal Mennell, Julie, Northumbria University, UK Mi, Bin, Boston Scientific Corporation, USA Minas, Graca, University of Minho, Portugal Moghavvemi, Mahmoud, University of Malaya, Malaysia Mohammadi, Mohammad-Reza, University of Cambridge, UK Molina Flores, Esteban, Benemérita Universidad Autónoma de Puebla, Mexico Moradi, Majid, University of Kerman, Iran Morello, Rosario, University "Mediterranea" of Reggio Calabria, Italy Mounir, Ben Ali, University of Sousse, Tunisia Mulla, Imtiaz Sirajuddin, National Chemical Laboratory, Pune, India Nabok, Aleksey, Sheffield Hallam University, UK Neelamegam, Periasamy, Sastra Deemed University, India Neshkova, Milka, Bulgarian Academy of Sciences, Bulgaria Oberhammer, Joachim, Royal Institute of Technology, Sweden Ould Lahoucine, Cherif, University of Guelma, Algeria Pamidighanta, Sayanu, Bharat Electronics Limited (BEL), India Pan, Jisheng, Institute of Materials Research & Engineering, Singapore Park, Joon-Shik, Korea Electronics Technology Institute, Korea South Penza, Michele, ENEA C.R., Italy Pereira, Jose Miguel, Instituto Politecnico de Setebal, Portugal Petsev, Dimiter, University of New Mexico, USA Pogacnik, Lea, University of Ljubljana, Slovenia Post, Michael, National Research Council, Canada Prance, Robert, University of Sussex, UK Prasad, Ambika, Gulbarga University, India Prateepasen, Asa, Kingmoungut's University of Technology, Thailand Pullini, Daniele, Centro Ricerche FIAT, Italy Pumera, Martin, National Institute for Materials Science, Japan Radhakrishnan, S. National Chemical Laboratory, Pune, India Rajanna, K., Indian Institute of Science, India Ramadan, Qasem, Institute of Microelectronics, Singapore Rao, Basuthkar, Tata Inst. of Fundamental Research, India Raoof, Kosai, Joseph Fourier University of Grenoble, France Reig, Candid, University of Valencia, Spain Restivo, Maria Teresa, University of Porto, Portugal Robert, Michel, University Henri Poincare, France Rezazadeh, Ghader, Urmia University, Iran Royo, Santiago, Universitat Politecnica de Catalunya, Spain Rodriguez, Angel, Universidad Politecnica de Cataluna, Spain Rothberg, Steve, Loughborough University, UK Sadana, Ajit, University of Mississippi, USA Sadeghian Marnani, Hamed, TU Delft, The Netherlands Sandacci, Serghei, Sensor Technology Ltd., UK Schneider, John K., Ultra-Scan Corporation, USA Sengupta, Deepak, Advance Bio-Photonics, India Shah, Kriyang, La Trobe University, Australia Sapozhnikova, Ksenia, D.I.Mendeleyev Institute for Metrology, Russia Saxena, Vibha, Bhbha Atomic Research Centre, Mumbai, India Seif, Selemani, Alabama A & M University, USA Seifter, Achim, Los Alamos National Laboratory, USA Silva Girao, Pedro, Technical University of Lisbon, Portugal Singh, V. R., National Physical Laboratory, India Slomovitz, Daniel, UTE, Uruguay Smith, Martin, Open University, UK Soleymanpour, Ahmad, Damghan Basic Science University, Iran Somani, Prakash R., Centre for Materials for Electronics Technol., India Srinivas, Talabattula, Indian Institute of Science, Bangalore, India Srivastava, Arvind K., NanoSonix Inc., USA Stefan-van Staden, Raluca-Ioana, University of Pretoria, South Africa Sumriddetchka, Sarun, National Electronics and Computer Technology Center, Thailand Sun, Chengliang, Polytechnic University, Hong-Kong Sun, Dongming, Jilin University, China Sun, Junhua, Beijing University of Aeronautics and Astronautics, China Sun, Zhiqiang, Central South University, China Suri, C. Raman, Institute of Microbial Technology, India Sysoev, Victor, Saratov State Technical University, Russia Szewczyk, Roman, Industrial Research Inst. for Automation and Measurement, Poland Tan, Ooi Kiang, Nanyang Technological University, Singapore, Tang, Dianping, Southwest University, China Tang, Jaw-Luen, National Chung Cheng University, Taiwan Teker, Kasif, Frostburg State University, USA Thirunavukkarasu, I., Manipal University Karnataka, India Thumbavanam Pad, Kartik, Carnegie Mellon University, USA Tian, Gui Yun, University of Newcastle, UK Tsiantos, Vassilios, Technological Educational Institute of Kaval, Greece Tsigara, Anna, National Hellenic Research Foundation, Greece Twomey, Karen, University College Cork, Ireland Valente, Antonio, University, Vila Real, - U.T.A.D., Portugal Vanga, Raghav Rao, Summit Technology Services, Inc., USA Vaseashta, Ashok, Marshall University, USA Vazquez, Carmen, Carlos III University in Madrid, Spain Vieira, Manuela, Instituto Superior de Engenharia de Lisboa, Portugal Vigna, Benedetto, STMicroelectronics, Italy Vrba, Radimir, Brno University of Technology, Czech Republic Wandelt, Barbara, Technical University of Lodz, Poland Wang, Jiangping, Xi'an Shiyou University, China Wang, Kedong, Beihang University, China Wang, Liang, Pacific Northwest National Laboratory, USA Wang, Mi, University of Leeds, UK Wang, Shinn-Fwu, Ching Yun University, Taiwan Wang, Wei-Chih, University of Washington, USA Wang, Wensheng, University of Pennsylvania, USA Watson, Steven, Center for NanoSpace Technologies Inc., USA Weiping, Yan, Dalian University of Technology, China Wells, Stephen, Southern Company Services, USA Wolkenberg, Andrzej, Institute of Electron Technology, Poland Woods, R. Clive, Louisiana State University, USA Wu, DerHo, National Pingtung Univ. of Science and Technology, Taiwan Wu, Zhaoyang, Hunan University, China Xiu Tao, Ge, Chuzhou University, China Xu, Lisheng, The Chinese University of Hong Kong, Hong Kong Xu, Sen, Drexel University, USA Xu, Tao, University of California, Irvine, USA Yang, Dongfang, National Research Council, Canada Yang, Shuang-Hua, Loughborough University, UK Yang, Wuqiang, The University of Manchester, UK Yang, Xiaoling, University of Georgia, Athens, GA, USA Yaping Dan, Harvard University, USA Ymeti, Aurel, University of Twente, Netherland Yong Zhao, Northeastern University, China Yu, Haihu, Wuhan University of Technology, China Yuan, Yong, Massey University, New Zealand Yufera Garcia, Alberto, Seville University, Spain Zakaria, Zulkarnay, University Malaysia Perlis, Malaysia Zagnoni, Michele, University of Southampton, UK Zamani, Cyrus, Universitat de Barcelona, Spain Zeni, Luigi, Second University of Naples, Italy Zhang, Minglong, Shanghai University, China Zhang, Qintao, University of California at Berkeley, USA Zhang, Weiping, Shanghai Jiao Tong University, China Zhang, Wenming, Shanghai Jiao Tong University, China Zhang, Xueji, World Precision Instruments, Inc., USA Zhong, Haoxiang, Henan Normal University, China Zhu, Qing, Fujifilm Dimatix, Inc., USA Zorzano, Luis, Universidad de La Rioja, Spain Zourob, Mohammed, University of Cambridge, UK Sensors & Transducers Journal (ISSN 1726-5479) is a peer review international journal published monthly online by International Frequency Sensor Association (IFSA). Available in electronic and on CD. Copyright © 2011 by International Frequency Sensor Association. All rights reserved. SSeennssoorrss && TTrraannssdduucceerrss JJoouurrnnaall CCoonntteennttss Volume 126 Issue 3 March 2011 www.sensorsportal.com ISSN 1726-5479 Research Articles Human Heat Generator for Energy Scavenging with Wearable Thermopiles Vladimir Leonov.................................................................................................................................. 1 Fabrication of a Highly-sensitive Acetylcholine Sensor Based on AChOx Immobilized Smart-chips M. M. Rahman .................................................................................................................................... 11 Development of Pressure-Temperature Integrated Multifunction Sensor Using Piezo- Resistive Element Palash K. Kundu, Gautam Sarkar and Chandan Dutta...................................................................... 19 Prototype for Managing the Wheelchair Movements by Accelerometry Daniel Alves Fusco, Alexandre Balbinot ............................................................................................ 31 Design of an Automatic Path Finding Wheelchair with Intelligent Guidance System Apratim Majumder, Niladri Banerjee, Shikha Nayak and B. Chakraborty. .................................................................................. 42 Design and Manufacture an Ultrasonic Dispersion System with Automatic Frequency Adjusting Property Herlina Abdul Rahim, Javad Abbaszadeh Bargoshadi, Sahar Sarrafi Ruzairi Abdul Rahim ............. 52 A Fusion Approach to Feature Extraction by Wavelet Decomposition and Principal Component Analysis in Transient Signal Processing of SAW Odor Sensor Array Prashant Singh and R. D. S. Yadava. ................................................................................................ 64 Development of a Simple Traffic Sensor and System with Vehicle Classification Based on PVDF Film Element D. R. Santoso, Abdurrouf, L. Nurriyah................................................................................................ 74 Practical Investigation of an Acoustic Encoder Mohammad A. Alia, Mohammad Al-Khedher, Mazouz Salahat ......................................................... 85 A Novel Method to Balance Inverted Pendulum by Angle Sensing Using Fuzzy Logic Supervised PID Controller Sanjeev Kumar, Ashutosh K. Agarwal, Arpita Gupta, Himanshu Tripathi, Prachi Mohan Kulshrestha ................................................................................................................. 92 An Integrated Expert Controller for the Oven Temperature Control System Nagabhushana Katte, Nagabhushan Raju Konduru, Bhaskar Pobbathi, and Parvathi Sidaraddi..... 101 Implementation of PID Controller in MATLAB for Real Time DC Motor Speed Control System Manjunatha Reddy H. K., Immanuel J., Parvathi C. S., P. Bhaskar and L. S. Sudheer .................... 110 Applying Time Series Analysis Model to Temperature Data in Greenhouses Abdelhafid Hasni, Zouaoui Chikr-el-Mezouar, Belkacem Draoui and Thierry Boulard ...................... 119 Growth Factor Inhibiting PKC Sensor in E-coli Environment Using Classification Technique and ANN Method T. K. Basak, T.Ramanujam, S. Jeybalan, Madhubala Bhatt, Deepali Garg, Richa Garg .................. 125 Authors are encouraged to submit article in MS Word (doc) and Acrobat (pdf) formats by e-mail: editor@sensorsportal.com Please visit journal’s webpage with preparation instructions: http://www.sensorsportal.com/HTML/DIGEST/Submition.htm International Frequency Sensor Association (IFSA). https://www.lesensor.com/sensor/Profiles/CreateNewAccount.aspx?sensor_portal=ls10001 Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 1-10 1 SSSeeennnsssooorrrsss &&& TTTrrraaannnsssddduuuccceeerrrsss ISSN 1726-5479 © 2011 by IFSA http://www.sensorsportal.com Human Heat Generator for Energy Scavenging with Wearable Thermopiles Vladimir Leonov Imec, Kapeldreef 75, 3001 Leuven, Belgium, Tel.: +32-16-288-367, Fax: +32-16-288-500 E-mail: leonov@imec.be Received: 28 February 2011 /Accepted: 17 March 2011 /Published: 29 March 2011 Abstract: Human beings and other warmblooded animals and birds constantly generate heat. A heat flow of one-to-several watt can be observed through a thermoelectric generator (TEG) worn by a person. The TEG would convert natural heat flow rejected from the body into electrical power of the order of milliwatts. Such a TEG can outperform a battery of the same weight in several months of use. Therefore, it could be a successful competitor to a battery as a power supply for low-power wearable sensors. Such a green power source could be attractive for the market because it requires no technical service for the entire service life of device, and can be safely disposed at the end of its life. To correctly perform the design optimization of wearable TEG, the knowledge of human body properties is important. This paper discusses thermal properties of human beings relevant to designing of wearable TEG. Copyright © 2011 IFSA. Keywords: The human being, Thermal resistance of human body, Thermoelectric generator, Energy scavenging, Wearable device. 1. Introduction Several self-powered wireless sensors for health monitoring have been recently demonstrated [1-3]. These low-power wearable sensors are fully powered by energy scavengers, mainly by a thermoelectric generator (TEG), but partially by photovoltaic (PV) cells either. The preferable locations for a wearable device are the wrist, where the device could resemble a watch or be combined with a watch, and thin devices in clothing. Powering a device in a piece of clothing by PV cells is not an easy task because most of people spend their daytime indoors. In this case, not much power can be scavenged. Low-power wireless sensors, depending on the data transmission rate and range, consume Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 1-10 2 from 10 μW [4] to about 1 mW minimum. To provide such power indoors with PV cells, a large area of clothing must be covered by the cells. In addition, any piece of clothing worn on top of PV cells obscures them from light. Therefore, the advantage of using PV cells outdoors is frequently lost. Of course, transparent windows for light could be made in those pieces of clothing, but this solution would be hardy accepted on the market. A TEG does not need to be on the outer surface of textile. If it is covered by a piece of clothing, it can still produce enough power for a sensor. It works at night, too. Therefore, a thermoelectrically-powered device in garments seems more practical on the market than the one with PV cells. When discussing energy scavenging, the comparison with a battery is necessary to define usefulness or useless of using the harvester in particular applications. While on heated machinery a TEG quickly outperforms any existing battery, it is not obvious that the same situation could happen in a wearable device. This is because the temperature difference between the human skin and environment is low. Let us estimate capabilities of both, an alkaline AA battery and a thermoelectric energy scavenger, both used as alternative solutions for powering a wearable sensor. Modern batteries could reliably power a wearable device for several years. They could be encapsulated to sustain machine washing. However, replacement of such a battery could be a complex or impossible task, especially for elderly people who require most of healthcare. Therefore, the battery must provide power either for the entire service life of the device, or of the piece of clothing. We assume that for the energy saving reason, the electronic module functions at low voltage of 1 V, but it accepts a voltage of up to 1.5 V with proportional increase of power consumption. One AA battery has 3 000 mA hr and provides a voltage of 1.5 V in the beginning, but about 0.8 V at the end of its service life. We account only for the capacity of the battery provided at a voltage not less than 1 V, and subtract 10 % of the capacity from 3 000 mA hr. Therefore, our battery has an initial capacity of 2 700 mA hr. It weighs 24 g and has a volume of 8.4 cm3. Its average discharge voltage exceeds the minimum voltage and is assumed to be 1.25 V. To compare it with a TEG, we assume that the latter has a device density of 0.8 g/cm3, i.e., about 3.5 times less than the density of the battery. This density corresponds to the best currently available wearable TEGs, because they are semi-empty and mainly filled with air. We assume that the optimized TEG of the same weight as a battery has a thickness of 1 cm, so that it occupies 30 cm2 on the subject’s body. The power generation in such a TEG could be within the 10 to 20 μW/cm2 on average depending on particular location of the TEG. We will use a moderate value of 15 μW/cm2 at 1 V output. We assume that the sensor node monitors the subject’s heath for 24 hrs a day. At night, the TEG on his/her body will still produce power, but less than at daytime. For simplicity we will not account for this power as the worst-case scenario. We assume that when the subject is not in a bed, the TEG effectively generates power, i.e., it produces power 16 hrs per day. At a power production of 15 μW/cm2, the TEG produces 0.45 mW. Therefore, it supplies 16 × 0.45 = 7.2 mW hr per day. This power is uniformly redistributed over 24 hr, therefore, a power consumption of the sensor node cannot exceed 7.2 / 24 = 0.3 mW at 0.3 mA and 1 V, on average. This TEG will supply power to the sensor node for unlimited time. The battery is however limited by its shelf life and capacity. The average power consumption of a sensor node powered by above battery will be 0.3 × 1.25 = 0.375 mW provided at 0.3 mA. According to the specifications from Duracell, from 4% to 7% of the initial charge is lost during the first year if stored at room temperature. The battery self-discharge is accelerated at elevated temperatures such as a skin temperature of 35-36 °C. Assuming 7% self-discharge, which is not the worst-case scenario, by the end of the first year, the battery capacity will be 2 580 mA hr. (Machine washing is typically performed at higher temperatures, so battery self-discharge would be very much accelerated at that period of time.) The sensor node will consume this energy during 2580/(0.3×24)=358 days, i.e., in less than one year. From this analysis, we conclude that a wearable TEG outperforms an AA battery in about 1 year of use. The reader can argue that Li AA batteries have a weight of 14.5 g and have more uniform voltage output. However, on one hand, an assumed average density of 0.8 g/cm3 in a TEG refers to already fabricated demonstrators. It is expected to be essentially decreased in the future by, e.g., a factor of two Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 1-10 3 to four. On the other hand, AA battery in clothing is not comfortable. Thinner batteries have worse capacity per unit volume and weight. Therefore, the comparison of a TEG and battery performed above looks quite reasonable and adequate. Lithium batteries also provide higher average discharge voltage of about 1.45 V, according specifications from Energizer. This, in turn, further increases power consumption and decreases the battery service life. The latter could be however avoided by voltage down-conversion to 1 V assumed above (but with the related power loss). Environmental impact is also important. A TEG in garments can be safely disposed, while an integrated “Li battery waste” in clothes is hardly to be safely disposed and could create an environmental problem. Depending on the design of wearable TEG, its location, particular application and the user-related scenario of its use, the situation however may turn in favor of a battery. For example, the person may have several changeable pieces of clothing with the same type of sensors integrated in the garment. We now assume that 5 pieces of clothing are worn by the user sequentially, i.e., one is worn, and the other four are washed in a washing machine or placed in a wardrobe. In such scenario, each of 5 TEGs produces 5 times less power, because it is worn only once per 5 days on average. The TEG therefore generates only 1.44 mW hr per day. In such user scenario, the system stays in a standby regime most of the time. Assuming a standby power of 1 μW, practically all produced power, i.e., 1.42 mW hr per day is still used in active regime. The AA alkaline battery at such power consumption will last 5 years. After 5 years of use, or earlier, the piece of clothing most probably will reach the end of life and will be replaced. In this scenario, the battery could probably produce more power than the TEG. The environmental aspects, however, may still dominate. Furthermore, the production cost of a TEG can be lower than that of a battery, especially, the lithium one. Such a TEG can be safer, for example, it will not explode in case that the device in clothes is overheated by an iron during its pressing. These factors can be dominant on the market. Therefore, even in such user scenario, at equal power produced by a TEG and the battery, there are still certain advantages that could be offered by a TEG. However, in order to be competitive to a battery, the TEG must be attentively designed so as to reach competitive power generation used in above comparisons. If the TEG produces much less than 10-15 μW/cm2 (or about 15 μW/g), its competitiveness to both a battery and PV cells could become questionable. Therefore, further significant reduction of the TEG density, at the same power produced per its unit weight, is important for the market success of body-powered devices in garments. On the same reason, a wearable TEG must be optimized to outperform a battery in several months of use, and knowledge of human body properties is very important for the optimization. 2. The Human Body and Thermoelectric Generator Discussion on the thermal properties of human body and their importance for designing a wearable TEG was presented in [5, 6]. Analysis of the thermal circuit of a TEG, Fig. 1, shows that heat flow, W, to large extent depends on the thermal resistance of human being: )RRR(/TW TEGairbody  , (1) where ∆T is the temperature difference between the heat generating zone of a heat generator (i.e., where the heat is produced) and the ambient, Rbody, Rair and RTEG are thermal resistances of the human being, ambient air and the TEG, respectively; note that Rbody and Rair include interfaces. The temperature difference that appears on a thermopile, ∆Ttp, is: )RRR(/RTT TEGairbodyTEGtp  . (2) It is obvious from Eq. (1) that if the TEG has extremely high thermal resistance, heat flow approaches zero. If, in turn, it has extremely low thermal resistance, ∆Ttp turns to zero, Eq. (2). It is known that the Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 1-10 4 power produced by a thermopile is a product of the heat flow through the thermopile and its conversion efficiency. The latter is proportional to ∆Ttp, therefore the power is proportional to the product of the heat flow and ∆Ttp. If one of two factors approaches zero, power minimizes. The maximum power, Pmax, is reached at: )RR(/TT Z P airbodyopt,tpmax  8 , (3) where Z is the thermoelectric figure-of-merit, Z = S2σ/k, S is the Seebeck coefficient, σ and k are electrical and thermal conductivities, respectively, ∆Ttp,opt is the optimal temperature difference on the thermopile, ∆Ttp,opt = ∆T/[2(1+1/N)], N is the thermal insulation factor, N ≈ Rins / (Rbody + Rair), and Rins is the parasitic thermal resistance between the hot and cold plates. If the thermal resistance of insulating supports/encapsulation between the plates is less than the thermal resistance of the thermal generator, the wearable TEG becomes inefficient irrespective of thermoelectric materials used, and of the shape and quantity of thermocouple legs. Therefore, N ≥ 1 must be provided on the design stage, and the thermal resistance of human being must be known for the TEG optimization. Fig. 1. A TEG on a person, one of possible designs of such TEG, and its thermal circuit. The TEG is the thermal load of natural thermal generator. Thermally insulating encapsulation together with air inside the TEG is responsible for parasitic thermal resistance Rins that is observed in parallel to the thermopile. 3. The Thermal Resistance of Human Body The thermal resistance of humans, Rbody, can be calculated as (Tcore – Tskin)/W. The core temperature can be evaluated from tympanic, esophageal, oral or rectal temperature, but typically it stays around Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 1-10 5 37°C. The skin temperature is variable and depends on many factors. It cannot be assumed and must be measured. At the same moment, the heat flow can be measured by using a thermopile with known thermal resistance attached to the skin. Typically, off-the-shelf thermopiles have relatively low thermal resistance and the heat flow according to Eq. (1) does not essentially change. This paper is devoted to studying human body properties in case of a TEG attached to the skin. Such TEG must have the thermal resistance close to the one of thermal generator, Eq. (2), otherwise, the temperature drop on the thermopile and the power will be too low. As a result, the heat flow through the optimized TEG must be less than naturally observed heat flow on the skin. To obtain results on the body’s thermal resistance relevant to the case of near-optimum TEG, the latter must have a thermal resistance of the order of several hundred cm2K/W. For experiments, the TEG shown in Fig. 1 was fabricated where a thermopile was mounted between two aluminum plates. The circular hot plate has an area of 7 cm2. A rectangular cold plate attached to the thermopile was cooled in the experiment to different temperatures. The resulting open-circuit Seebeck voltage was proportional to the heat flow. A K-type thermocouple has been glued to the hot plate from the skin side for monitoring the skin temperature. The TEG is shown in Fig. 2 (a). Similar TEG modules have recently been used in a self-powered electrocardiography system-in-a-shirt [3]. In the first experiment, the TEG was sequentially attached to the wrist and to the leg using elastic band. It is clear that the TEG must be attached tightly to skin for good thermal contact. However, too tight device is uncomfortable. The criterion used in the experiment was sensation of comfort. The cold plate was cooled to different extent and each time, before recording any experimental data, a steady thermal equilibrium with the body was reached so that both the skin temperature and Seebeck voltage seem to stabilize. The experiments have been conducted indoors, in the office, at temperatures within the 21.0 °C to 23.7 °C range, with no air conditioning, on a sitting person, in a course of several days. In the experiment, the radial artery allowed high heat flow. However, the skin temperature showed very weak dependence on heat flow and decreased to 30 °C only at a heat flow of about 1 W through the TEG, Fig. 2 (b). The maximum heat flow reached in the leg was lower, but the skin temperature rapidly dropped below 20°C. This difference observed between the wrist and leg is due to the difference in thermal resistance of these body parts. 10 20 30 40 0 20 40 60 80 100 120 140 160 S ki n te m p e ra tu re (° C ) Heat flow (mW/cm2) 1 2 a b Fig. 2. (a) The TEG for the measurement of heat flow on a person, and (b) the dependence of skin temperature under this TEG on heat flow: (1) on the left wrist, on the radial artery, and (2) on the front side of right leg, approximately 25 cm above the knee. Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 1-10 6 The thermal resistance extracted from the measurement results is shown in Fig. 3. As has been seen in the experiment, at high heat flows, i.e., at those essentially exceeding natural one, slow drifting of the temperature of large part of the leg during an hour. Therefore, to minimize this adverse effect of body temperature variation on accuracy of measurements, another experiment has been conducted with halved contact area of the TEG with the body, namely, on 1.6×2.1 cm2 area. Fig. 3. Dependence of the thermal resistance of the studied subject on heat flow measured in the office. The TEG is located: (1) on the left wrist, on the radial artery, and (2) on the front side of right leg, approximately 25 cm above the knee. Open data points correspond to 7 cm2 hot plate area, closed points are measured on the hot plate area of 3.4 cm2. The results shown in Fig. 3 indicate that at high heat flows the thermal resistance stabilizes. It is 200 cm2K/W in the leg and 60 cm2K/W in the wrist. The two measured locations reflect two very different conditions on the human body. On the leg, a thick layer of muscles results in much higher thermal resistance than in the location over an artery. The muscles are the dominant natural thermal insulator in the human body. In contrary, arteries are effective heat spreaders and heat exchangers. At indoor temperatures, the temperature of the arterial blood is close to 37°C. The distance from the artery in the wrist to the skin is only a few millimeters. Therefore, much higher heat flow can be obtained through the TEG located in proximity to an artery, and the thermal resistance reaches its minimum. Natural heat flow on the person’s skin depends on his/her body heat content. The latter depends on the physical activity (metabolic rate and speed of moving), clothes and ambient conditions (temperature, humidity, and wind). Therefore, the measured heat flow and thermal resistance very much depend on conditions of measurements. It is believed that in case of a person wearing appropriate clothes so as to maintain the body at thermal comfort (on the whole body level), most of the body area must show the thermal resistances between the two curves shown in Fig. 3. If the person is overheated and feels hot, the thermal resistance could go below the measured curves. If however the person feels cold, does not wear appropriate clothes in cold weather, or performs insufficient physical activity in cold environment to maintain his/her body temperature at the same level as in the office, the thermal resistance can easily exceed the highest measured values. In cold environment, the thermal resistance increases first in extremities due to body temperature regulation. To preserve the body core organs from overcooling, the vasoconstriction decreases the diameter of arteries and the heat supplied to Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 1-10 7 extremities essentially decreases either. In addition, the vein and artery patterns in extremities very much coincide with each other. As a result, the countercurrent heat exchange between arteries and veins further decreases the temperature in feet and hands, and can dramatically increase their thermal resistance well beyond the thermal resistance of the ambient air. In such case, not the ambient air, but the body itself limits the heat flow. It is known from medical study that a heat flow of a few to about 15 mW/cm2 is typically observed indoors depending on particular location on the person. Fig. 3 shows that in such case, the thermal resistance of human body can approach the one of ambient air and may become equally important for thermal optimization of a TEG. Therefore, a wearable TEG cannot be optimized unless the local thermal resistance of the body is accounted for device optimization. The following two factors are important for designing wearable TEGs. Depending on the heat flow, the body’s thermal resistance varies. The higher heat flow, the lower is the thermal resistance. However if the ambient temperature decreases and heat flow increases, the thermal resistance does not necessarily decrease. In fact, it typically increases [7]. However, in those medical studies, the body thermoregulation actively participated in changing the heat flow and its pattern. The skin temperature also varied with the ambient temperature. In our research, the body is maintained at or near thermal comfort at different ambient temperatures by wearing appropriate clothes, see, e.g., Fig. 4. Therefore, the thermoregulation mechanisms are either not active or not that active. This causes completely different reaction of the human body on local enhancement of heat flow, and the relevant study was necessary. Two wearable TEGs have been used in this work to measure the dependence of body’s thermal resistance on ambient temperature. The first one was described above. It has a cold plate of 3 cm × 4 cm and a thickness of 6.5 mm that is near-optimum thickness for maximum power generation per unit volume of a TEG [8]. The second TEG has four-stage thermopiles mounted on the hot plate of 5 cm2 and supplied with a radiator of 3 cm × 3 cm × 1.5 cm size. The two TEGs have been either integrated in the garments, or worn on the wrist as shown in Fig. 4. The experiments have been conducted under condition of subject’s feeling of thermal comfort. It does not necessarily mean that at sub-zero temperatures, the skin temperature, especially in hands, remains the same as indoors. In the experiment, it was a little lower. However, the clothes worn were typical to the corresponding weather conditions, and the sensation of comfort was reported. Therefore, the obtained results reflect typical body properties at typical conditions that might be expected in practical applications. a b Fig. 4. (a) Two TEGs integrated in jeans at about 25 cm above knees, and (b) the TEG with a radiator attached to the wrist. Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 1-10 8 The power generated by a TEG depends on heat flow. At the same temperature, the TEG shown in Fig. 4 (b) produces more power than the TEG shown in Fig. 2 (a) because radiator decreases thermal resistance of ambient air and increases heat flow. On one hand, such a bulky TEG with radiator is not comfortable to wear. It offers more power, but in cold environment, increased heat flow may make it uncomfortable (too cold). On the other hand, a thin TEG with a cold plate instead of radiator may not be competitive to PV cells and batteries in hot weather. To see the difference in thermal resistance of human body, the experiments were conducted with both TEGs at once. The thermal resistance of the leg measured with two TEGs as depicted in Fig. 4 (a) is plotted in Fig. 5 versus ambient temperature. The results shown in Fig. 5 indicate that a radiator on the cold side of a TEG decreases the body’s thermal resistance. In the experiment, the thermal resistance of the leg was decreased by a factor of two. The other useful observation is that the thermal resistance decreases at lower ambient temperatures. However, it seems to stabilize below 10°C.   0 200 400 600 800 1000 1200 -10 0 10 20 30 T h e rm a l re si st a n ce (c m 2 K /W ) Ambient temperature (oC) 1 2 Fig. 5. Dependence of the thermal resistance in legs on ambient temperature measured by using (1) the TEG with a hot plate and (2) the TEG with a radiator. The experimental setup is shown in Fig. 4 (a). The lines are guides for the eye. Completely different results have been obtained with the same two TEGs attached to the left wrist, on the radial artery. The average thermal resistance of the wrist measured in the office by using the TEG with a radiator is 110 cm2K/W, i.e., it coincides with a preliminary result of 110±20 cm2K/W obtained earlier [9]. This thermal resistance increases linearly to about 140 cm2K/W at ambient temperature of 2 °C. The TEG with a cold plate at 25°C has shown body’s thermal resistance of 160 cm2K/W. The thermal resistance of the body, as was observed first in [10], depends on heat flow. Therefore, the lower the heat flow, the higher is the body’s thermal resistance under the same environmental conditions, and at the same clothes and body heat content. The design of wearable TEG, in particular, its radiator, affects the thermal resistance of human body due to redirection of heat flows inside the body. However, the change in body’s thermal resistance requires thermal redesign of a TEG in accordance with Eq. (2) and the equation of its thermal matching with the thermal generator [11]. This mutual influence of human being and a wearable TEG must be accounted for designing of the latter and for comfort of the former. The other important factor that must be accounted for is the maximum heat flow acceptable with no cold-related discomfort. The measurements show that the heat flow up to 15-25 mW/cm2 is acceptable Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 1-10 9 at indoor temperatures, depending on device location, and on the person under study. Larger heat flow is impossible on a standing or sitting person indoors because the required dimensions of radiator would be unacceptably large. In cold weather, sensation of cold caused by the wrist device has been reported only at the heat flow exceeding 100-130 mW/cm2. On the neck (also near the artery) the limit was 100 mW/cm2 at ambient temperature of 0 °C. On legs, a heat flow of 70 mW/cm2 was still quite acceptable at ambient temperature of –4 °C, but the acceptable limit of the heat flow has not been reached. On the trunk and head, the sensitivity to cold is higher. As reported earlier, a heat flow of 20 to 30 mW/cm2 on the head and chest may already cause (with certain probability) unwanted sensation of cold according to user’s responses/complaints [8, 9]. Some other experimental results on heat flows in wearable TEGs and on the thermal resistance of human body, which can be useful for designing wearable energy harvesters, can be found in the literature [8-13]. The TEG modules like the one shown in Fig. 2 (a) have been integrated into the office shirt for powering an autonomous wireless electrocardiography system [3]. Fourteen modules have been integrated in the most convenient locations that showed the lowest body’s thermal resistance either. The measurements have shown that the TEG was neither cold nor obtrusive to the user. In winter, the outdoor pieces of clothing will be worn on top of integrated or worn TEGs. However, as measured outdoors on a person wearing a thick jacket, no sensation of discomfort caused by enhanced heat flow through the TEG has been reported, at least down to an ambient temperature of –2 °C. The TEG with a radiator shown in Fig. 4 (b) and worn on the wrist demonstrated a power of up to 7 mW at –2°C. Even at such a low ambient temperature and the related high heat flow, this power has been generated with no sensation of cold, unobtrusively. However, this TEG seems to be too bulky for practical applications. Therefore, in an unobtrusive (thin) wearable TEG, the power will be lower. At proper positioning on the human body, the generated power will not be limited by the sensation of cold, but by the form factor of a TEG. 4. Conclusions Wearable devices consuming a power of 1-2 mW or less, including wireless sensors for health monitoring, can be powered by using human body heat. Thermoelectric conversion of human heat seems to be the best and the least obtrusive way to make wearable devices self-powered and with unlimited service life. Wearable thermoelectric generators are expected to outperform a battery of the same weight after less than one year of use. However, as shown in this paper, to be capable to compete with the battery, the TEG must be fully optimized. Its thermal optimization is strongly affected by thermal properties of human being, namely, by local thermal resistance of the body and location- dependent acceptable heat flows. It is found that the thermal resistance of human being depends on heat flow, and has linear dependence on heat flow indoors. The maximum power reached in the wearable TEG with a radiator is 1.4 mW/cm2 at ambient temperature of –2°C. In unobtrusive thermally optimized TEG, i.e., in a thin one, the power generation at typical indoor conditions should be limited by about 10-15 μW/cm2, or 20-25 μW/cm3, which is however enough to successively compete with heavy batteries. Thermally optimized TEG acts as a thermal isolation, i.e., similar to fabric. Therefore, no sensation of cold caused by wearable thermoelectric generators is expected. Acknowledgements This work has been performed within the Imec’s Human++ program on smart wireless sensors. Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 1-10 10 References [1]. T. Torfs, V. Leonov, R. J. M. Vullers, Pulse oximeter fully powered by human body heat, Sensors & Transducers J., Vol. 80, Issue 6, 2007, pp. 1230-1238, http://www.sensorsportal.com/HTML/DIGEST/P_151.htm [2]. M. Van Bavel, V. Leonov, R. F. Yazicioglu, T. Torfs, C. Van Hoof, N. Posthuma, R. J. M. Vullers, Wearable Battery-Free Wireless 2-Channel EEG Systems Powered by Energy Scavengers, Sensors & Transducers J., Vol. 94, Issue 7, 2008, pp. 103-115, http://www.sensorsportal.com/HTML/DIGEST/P_300.htm [3]. V. Leonov, T. Torfs, R. J. M. Vullers, C. Van Hoof, Smart wireless sensors integrated in clothing: an electrocardiography system in a shirt powered using human body heat, Sensors & Transducers J., Vol. 107, Issue 8, 2009, pp. 165-176, http://www.sensorsportal.com/HTML/DIGEST/P_485.htm [4]. V. Pop, J. van de Molengraft, F. Schnitzler, J. Penders, R. van Schaijk, R. Vullers, Power Optimization for Wireless Autonomous Transducer Solutions, in Proc. PowerMEMS 2008 and MicroEMS 2008 Workshop, Sendai, Japan, 9-12 November 2008, pp. 141-144. [5]. V. Leonov, Z. Wang, P. Fiorini, C. Van Hoof, Modeling of micromachined thermopiles powered from the human body for energy harvesting in wearable devices, Sensors & Transducers J., Vol. 109, Issue 4, 2009, pp. 29-43. http://www.sensorsportal.com/HTML/DIGEST/P_410.htm [6]. V. Leonov, R. J. M. Vullers, Wearable thermoelectric generators for body-powered devices, J. of Electronic Materials, Vol. 38, Issue 7, 2009, pp. 1491-1498. [7]. Heat flow from animals and man, J. Monteith, L. Mount, Eds., Butterworths, London, 1974. [8]. V. Leonov, Theoretical performance characteristics of wearable thermoelectric generators, Advances in Science and Technology, Vol. 74, 2010, pp. 9-14, Trans Tech Publications: Switzerland (Proc. 12th Int. Conf. on Modern Materials and Technologies (CIMTEC 2010), Montecatini Terme, Italy, 6-18 June 2010. [9]. V. Leonov, Z. Wang, R. Pellens, C. Gui, R. Vullers, J. Su, Simulations of a non-planar lithography and of performance characteristics of arcade microthermopiles for energy scavenging, in Proc. of the 5th Int. Energy Conversion Engineering Conf. (IECEC), St. Louis, MO, 25-27 June 2007. [10]. V. Leonov, T. Torfs, P. Fiorini, C. Van Hoof, Thermoelectric converters of human warmth for self- powered wireless sensor nodes, IEEE Sensors J., Vol. 7, Issue 5, 2007, pp. 650-657. [11]. V. Leonov, C. Van Hoof, R. J. M. Vullers, Thermoelectric and hybrid generators in wearable devices and clothes, in Proc. of the 6th Int. Workshop on Wearable and Implantable Body Sensor Networks (BSN’09), Berkeley, CA, 3-5 June 2009, pp. 195-200. [12]. V. Leonov, Heat generator in humans and its interaction with wearable thermoelectric energy scavenger, in Proc. of the 10th Int. Workshop Micro and Nanotechnology for Power Generation and Energy Conversion Applications (PowerMEMS), Tech. Digest: Oral sessions, Leuven, Belgium, 30 November – 3 December 2010, pp. 231-234. [13]. V. Leonov, Energy harvesting for self-powered wearable devices. In: Wearable monitoring systems, A. Bonfiglio, D. De Rossi, Eds., Springer, New York, USA, 2011, pp. 27-49. ___________________ 2011 Copyright ©, International Frequency Sensor Association (IFSA). All rights reserved. (http://www.sensorsportal.com) Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 11-18 11 SSSeeennnsssooorrrsss &&& TTTrrraaannnsssddduuuccceeerrrsss ISSN 1726-5479 © 2011 by IFSA http://www.sensorsportal.com Fabrication of a Highly-sensitive Acetylcholine Sensor Based on AChOx Immobilized Smart-chips M. M. RAHMAN Venture Business Laboratory, Department of Electrical and Electronic Engineering, Toyohashi University of Technology, 1-1 Hibarigaoka, Tempeku-cho, Toyohashi, Aichi 441-8580, Japan Tel.: +81-532-44-6974 E-mail: mmrahmanh@gmail.com Received: 12 January 2011 /Accepted: 17 March 2011 /Published: 29 March 2011 Abstract: Acetylcholine (ACh) sensor based on acetylcholine oxidase (AChOx) on EDC activated thioglycolic acid self-assembled monolayer (TGA-SAM) using smart-chip has been developed. The simple cyclic voltammetry (CV, at 0.1 V/s) technique is performed in total investigation, where 0.5M K3Fe(CN)6 is utilized as a standard mediator in phosphate buffer solution (PBS, 0.1M). The ACh sensor exhibited a lower detection limit (DL, 0.1392 ± 0.005 nM), a wide linear dynamic range (LDR, 1.0 nM to 1.0 mM), good linearity (R=0.9951), and higher sensitivity (7.3543 ± 0.2 μAμM-1cm-2), and required small sample volume (70.0 μL) as well as good stability and reproducibility. The smart-chip system employed a simple and efficient approach to the immobilization of enzymes onto active sensitive surface, which can enhance sensor performances to a large group of bio-molecules for wide range of biomedical applications in health care fields. Copyright © 2011 IFSA. Keywords: Acetylcholine, Acetylcholine esterase, Cyclic voltammetry, Smart-chip, Mediator. 1. Introduction Acetylcholine (ACh), the first identified neurotransmitter, can be found in both the peripheral nervous system and the central nervous system in mammals, including humans [1]. Recently the monitoring with a neurotransmitter in the living thing is interesting, despite of a research approached last decades [2, 3]. The low concentration from which the neurotransmitter is released to the liquid outside the cell space in the neural network of the brain technically makes their detection difficult. In the peripheral nervous system, ACh binds to ACh-receptors and regulates muscle contraction. In the central nervous system, ACh is performed alone with the most important neurotransmitter in a nervous sympathized Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 11-18 12 network. In terminal neurons, the neurotransmitter modulates the potential of the synaptic membrane by changing the activity of the ionotropic receptors [4]. Thus, the ions can move through the ionic channels. ACh is originated in preganglionic and motor neurons in living system, which affects learning, memory, and muscle tone [5]. Various conditions might continue if ACh cannot be metabolized, and it accumulates with the nerve tissue. There is clinical evidences indicating that some neuropsychiatric disorders (e.g., Parkinson’s disease, Alzheimer’s disease and Myasthenia Gravis) are correlated with dysfunctional ACh regulation [6, 7]. It plays the vital role by the procedure of involving to the behavioral function (e.g., arousal, attention, learning and memory) in the central nervous system, therefore sensing of ACh concentration is significant interest. ACh is synthesized in neurons from choline by means of choline acetyl transferase and acetyl co-enzyme A [8]. In order to investigate illnesses and to build up medicines, it is significant to evaluate ACh concentration with simple, fast, economical, and precise methods. Several methods have been introduced for the recognition of neurotransmitters such as, biochemical investigation using radioactive labels [9, 10], HPLC analysis [11, 12] and microdialysis with electrochemical detection [13, 14]. The key drawbacks of these methods are their poor chronological and spatial resolutions and the difficulty of the supplementary technical arrangement. Utilization of implantable micro-biosensors could overcome these difficulties with carbon-fiber-based electrodes looking particularly promising [15-18]. General techniques for ACh recognition comprise HPLC on microdialysis samples where an immobilized enzyme column and combined electrochemical detectors are used [19, 20]. A recent study used a pH sensitive poly(vinylchloride) membrane with a plasma-polymerized film as a potentiometric biosensor for ACh sensing; the resolution of this biosensor was 2.0 mM and the detection reached 100.0 uM. The performances of this device did not exceed those of ACh microsensors. However, for this device, the characteristic curve was not linear and calibration was required. There are a lot of applications for Enzyme Field-Effect Transistors glucose [21, 22], urea [23, 24], acetylcholine [25, 26], and alcohol [27], using different kinds of enzyme. The technique for immobilization is very essential, and the measurement background will concern sensitivity and stability. Sensors based on the principle of field effect in semiconductor structures have been comprehensively premeditated in the recent years [28- 30]. Recently, it is successfully developed a charge-transfer-type pH sensor based on a charge-coupled devices [31, 32]. In this article, it is proposed and demonstrated experimentally ACh sensors using smart -chips. Finally, it is developed ACh a highly-sensitive and low-detective ACh sensor using a smart-chip, which is designed and fabricated successfully using smart-chips. The most important characteristics of the developed ACh biosensor are higher sensitivity, large linear dynamic range, very low detection limit, small sample volume, highly stable, and reproducible. 2. Experimental 2.1. Chemical Reagents ACh, AChOx, TGA, and N-(3-Dimethylaminopropyl)-N'-ethylcarbodiimide hydrochloride (EDC) were purchased from Sigma-Aldrich company. All other chemicals were analytical grade and used without further purification. 0.1M PBS (pH ~7.3) is made by mixing the proper proportion of 0.2M NaH2PO4 and 0.2M Na2HPO4. All solutions were prepared with distilled water, which was obtained from a water purifying apparatus (12.0 M.Ω.cm) (AQUA MEDIA). 2.2. Apparatus The electrochemical experiments were performed using a votammetric analyzer (CV-50W, BAS). All investigations were carried out on electrochemical smart-chip (5mm5mm), which sensing area is 0.0805 cm2. The total investigations were carried out with enzyme modified chips composed as working, Pt layer as counter, and an Ag/AgCl (saturated KCl) as a reference electrode. CVs were Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 11-18 13 recorded at AChOx/TGA/AuE electrode from -0.1 to +0.5 V (versus Ag/AgCl) in a 0.1M PBS (pH 7.3) at 0.1V/s scan rate. 2.3. Construction of Smart-chips Electrochemical chips were fabricated by conventional photolithographic technique. Electrodes and passivation layers are developed on silicon wafer followed by dicing and packaging. N-doped Si wafers are prepared and overflowed by extra-pure water. In this step, all contaminations on the surface and native SiO2 layer are removed. At first, the wet oxidation is processed, and then dry oxidation is executed. Wafers are annealed in the condition of nitrogen. Aluminum is sputtered with Al-1% Si target. Then the photolithograph processes are applied. Resist coating, baking, exposure, and development are done by Kanto chemicals, and then it is rinsed by ionic water. Al is etched by etching solution. Resist is removed by plasma etching instrument. Then wafers are cleaned by acetone, methanol, and finally by plasma simultaneously. SiN layer is deposited by chemical vapor deposition. Surface of pad electrodes are etched by reactive ion etching. Finally residual resist layer is removed by plasma acing. After photolithographic process, Pt is sputtered by SP150-HTS. Then it is patterned by lift-off technique, in which wafers are immersed into the remover, and then washed with IPA. Photolithographic process is again investigated, where Ti is sputtered as a binding layer, and then Au is evaporated by deposition. Finally, Au layer is patterned by lift-off process. Palylene passivation layer is formed for the protection of the chip from water. Photolithographic process is executed again for pad protection. Then palylene-dimer is evaporated by deposition apparatus. Photolithography process is done again for patterning. Palylene layer is patterned by etching. Finally, unnecessary resists are removed by acetone and then wafer is cleaned by IPA. Resist is coated on a whole surface of the wafer for protection during dicing process. Si wafer is diced into pieces by dicing apparatus and stored into the desiccators. Resist on chip surface is removed by acetone and cleaned with IPA. The backside of the chip is roughed by a sheet of sandpaper for better adhesion and electrical stability. The chip is bonded with die and packaged by silver paste. It is dried in a drying oven. Pads on chip are connected to the package through gold wire with bonding machine. Finally, Si-based adhesive is put on the periphery of the chip to protect pads and gold wire from sample solution. Adhesive is dried at room temperature for 24 hours. The composition and thickness of each fabricated layer into micro-devices are mentioned in Table 1. Table 1. Function and thickness of every layer on the chip. Material Function Thickness (µm) Si Wafer material 500.00 SiO2 Insulation 0.40 Al Electric wiring 1.00 SiN Protection/separation 1.00 Ti/TiN Binding 0.15 Ti Binding 0.10 Pt Counter Electrode 0.25 Au Working Electrode 0.30 Palylene Passivation/protection 1.00 The semiconductor smart-chips were made on silicon wafer. Al was sputtered to fabricate as wiring and bonding pads. Pt/Ti/TiN was sputtered on thermal oxide of silicon and patterned by photolithography to fabricate counter electrode (CE). Ti/TiN layers were used for strong adhesion. Au/Ti were sputtered and lithographed, which made circular working electrode (WE) with a diameter of 1.6 mm in the center of the chip. After electrodes fabrication, palylene layer was fabricated by Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 11-18 14 evaporation method as a passivation layer. The wafer was diced to 5.0 mm square chips. This chip was bonded to a package by silver paste. Aluminum pads were connected to the package by gold wire. Finally, adhesive (Araldite, Hantsman, Japan) was put on the periphery of the chip, which prevents target solution from contacting pads, which is revealed in Fig. 1A. The magnified construction view of internal chip-center (sensing area) is presented in the Fig. 1B. A cross section of the sensor chip is shown in Fig. 1C. Fig. 1. Schematic diagram of (A) camera-view from top, (B) magnified schematic view of chip sensing-area, and (C) cross-sectional view of smart chips. 3. Results and Discussions Fig. 2 outlines the sensing protocol using the AChOx/TGA/Au-modified chip. It is used for covalent bond formation to immobilize the AChOx enzyme on the TGA-SAM via peptide conjugation in presence of activating agent (EDC). First, the self-assembled monolayer of TGA is formed by dropping the TGA solution onto the sensing area of chips for 2 hours. Then AChOx enzyme is immobilized on the TGA-SAM by the amide bond formation between the terminal-unbound carboxylic acids groups on the TGA film and the amine groups of the AChOx enzyme. Fig. 2. Fabrication scheme of Acetylcholin sensor on smart-chips. Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 11-18 15 For the stable attachment of AChOx onto TGA-SAM, the chip was kept for 24 hours into the refrigerator at 4.0 oC. The enzymatic reactions involved in the bio-sensing system for the detection of ACh are as follows (see Fig. 2): Acetylcholine + O2  AChOx Choline + CH3COO- + H2O2 (1) H2O2  1/2O2+1/2H2O +2e- (2) Reaction (1) is depended on ACh concentration in the reaction medium. On the chip, ACh is oxidized to form choline, acetate ion, and H2O2. Then H2O2 is auto-dissociation to produce the current (Reaction 2). This current is directly proportional to the ACh concentration in the solution system. Successful fabrication of ACh sensor using AChOx on the sensing area of smart chip has been demonstrated based on TGA-SAMs. Conventional electrochemical method, CV is the most versatile electroanalytical technique for the study of bioactive materials and species, which is widely used in industrial applications and academic or biochemical research or R & D approaches. CV is also an important technique to evaluate the coating/blocking property of the monolayer-coated electrodes using diffusion controlled redox couples. Chip surface was cleaned by Piranha solution [H2SO4:H2O2 (3:1)] and washed with pure water, then dried sufficiently by nitrogen. TGA was dissolved in ethanol to make 10.0 mM solution. TGA solution was dropped on a sensing area of chip, and then kept wet for 2 hours at room temperature. Fig. 3 shows the CVs of un-modified and TGA-SAM modified chip electrodes in 5.0 mM K3Fe(CN)6 with 0.1 M PBS as the supporting electrolyte at 0.1 V/s potential scan rates. It can be seen from the Fig. 3A, that the bare chip electrode (black-curve) shows a reversible voltammogram for the redox couple indicating that the electron transfer reaction is completely diffusion controlled. In contrast, the absence of any peak formation in the CVs of the TGA monolayer- modified electrodes (blue-curve) shows the redox reaction is inhibited or totally blocked. The CVs for TGA indicated a good blocking behavior for the electron transfer reaction, which means that a highly ordered, compact monolayer is formed on the sensing surface of the bio-devices [33]. AChOx was immobilized onto the TGA-SAM surface by peptide conjugation. First 10.0 mM EDC in 0.1 M PBS was put onto the TGA-SAM chip and then it was kept at 4.0 oC in the refrigerator to activate carboxylic group of TGA for 24 hours. Then EDC-treated electrode was washed gently with 0.1 M PBS to remove excess EDC. Then AChOx solution was dropped on the sensing area of chip and incubated in the refrigerator at 4.0 oC for 24 hours. AChOx was successfully immobilized onto TGA- SAM via covalent bond, which was confirmed by the current change in Fig. 3B. It showed that the CVs recorded for the bare electrode (black curve), AChOx/TGA/Au (blue curve), and 0.1 mM ACh solution (pink curve) of fabricated chip in a 0.1M PBS (pH 7.4) at 0.1 V/s scan rates. According to the control experiment, no significant change was observed when the CV was recorded with the bare electrode for 0.1 mM ACh in PBS due to the absence of AChOx enzyme. A small current change was observed at approximately +0.21V versus Ag/AgCl for 0.1 mM ACh solution, and this was due to the current of enzymatic reaction with ACh in presence of AChOx on the sensing surface of the chip. The enzymatic current approximately +0.21 V was executed to be increased on the increasing ACh concentration in the PBS buffer solution. The experimental conditions affecting the performances (detection limit, sensitivity, and response time) of the sensor were optimized in term of pH and shown in Fig. 3C. The pH of the buffer shows a strong effect on the activity of the sensing layer on the chips. The effect of the pH of the buffer on the current alteration was studied over the pH range of 2.0 to 9.0. Fig. 5 shows the CV peak currents obtained at different pH values for 0.1 mM ACh in 0.1 M PBS system. The peak height increased from pH 3.5 to 7.3 and then decreased above pH 7.3. The peak current decrease above pH 7.3 might have been due to the poor AChOx enzyme activity at higher basic medium. Therefore, the pH of the PBS system was preset at 7.3 throughout the experiments. Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 11-18 16 Fig. 3. (A) CVs of 5.0 mM K3Fe(CN)6/PBS for bare (black curve) and TGA-SAM (blue curve) modified electrodes; (B) CVs recorded in 5.0 mM K3Fe(CN)6/PBS of bare chip (black curve), AChOx (blue curve) modified , and in presence of 0.1 mM ACh (pink curve) solution; (C) Effect of pH of AChOx/TGA/Au electrode in 0.1 mM ACh solution with the smart-chip. Scan rate: 0.1 V/s, RE: Ag/AgCl (saturated KCl), Supporting electrolytes: 0.1M PBS. Cyclic voltammetric study using redox mediator [K3Fe(CN)6] was achieved to confirm the detection of ACh concentration. 70.0 µL of each ACh solution with mediator was dropped on the sensing area of the chip and investigated the sensing oxidation current in manifestation of mediator. Fig. 4A shows a typical CV (current-voltage) plot for the addition of varying amounts of ACh in a 0.1M PBS (pH 7.3). The currents increased gradually with increasing the concentration of ACh (1.0 nM to 100.0 mM) to a stable and saturated value. The AChOx/TGA/Au modified chip electrode achieved 92 % of steady state currents with in 10 sec. The increase of oxidation current was observed. This was because the ACh oxidized in presence of AChOx, and the current change lead to the negative current increased. Fig. 4B shows the calibration plots for the ACh obtained from the current-voltage responses with fabricated chips. Under the optimized conditions, the steady-state currents showed a linear relationship with the ACh concentration in the range from 0.1 nM to 1.0 mM, which is shown in Fig. 4C. The linear dependence of the ACh concentration yielded with a correlation coefficient of 0.9951. The detection limit for ACh was executed to be approximately 0.1392±0.005 nM, based on signal to noise ratio (3N/S). The ACh sensor also exhibited higher sensitivity, which was calculated as 7.3543 ± 0.2 μA.μM-1.cm-2. The sensitivity is much higher than the previously reported acetylcholine sensors [32-37]. A series of successive measurements of ACh in 0.1M PBS yielded a good reproducibility signal at AChOx/TGA/Au sensor with a relative standard deviation (RSD) 3.8 %. The sensor-to-sensor and run-to-run reproducibility for 0.1 mM ACh detection were found to be 1.8 and 1.4% respectively. To examine the long-term storage stabilities, the response of the AChOx/TGA/Au sensor was examined with the respect to the storage time. After each experiment, the sensor was washed with the buffer solution and stored in 0.1M PBS at 4.0 oC until use. The long-term storage stability of the sensor was tested every 3 days. The sensitivity retained 92 % of initial sensitivity up to Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 11-18 17 1 month. After 1 month, the response gradually decreased, possibly due to the loss of the enzyme activity. The above results clearly executed that the smart chip ACh sensor can be used for a month without any significant loss in sensitivity. The selectivity (interference effect) of the AChOx/TGA/Au sensor was evaluated by CV in the presence of other electroactive compounds such as lactate, glucose, glutamate, and uric acid etc. No significant current response was observed when 0.1 mM lactate, glucose, uric acids, and glutamate were introduced into the 0.1M PBS buffer. But when 0.1mM ACh solution was added to the electrolyte solution, a clear oxidation response was executed, indicating the selective detection of ACh with the AChOx/TGA/Au sensor layer. Significantly, at this concentration level, lactate, glucose, uric acid, and glutamate have no interferon for 0.1 mM ACh detection. Thus, the selectivity of the AChOx/TGA/Au smart sensor is acceptable for ACh detection in the presence of the common interfering compounds in normal physiological levels. Fig. 4. ACh responses of (A) variation of ACh concentrations, (B) calibration curve, and (C) linearity of developed on smart-chip at room conditions. 4. Conclusions Successful fabrication of highly sensitive acetylcholine sensor based on the immobilization of AChOx on the 5.0 mm square sensor chip has been investigated using redox mediator. Sensor chips are constructed by photolithographic technique, which is possible to detect the ACh solution of tiny volume. The sensor exhibited high sensitivity, low-detection limit with satisfactory stability, a large linear dynamic range, and reproducibility. The simple fabrication method of the biosensor has many advantages such as ease of fabrication, enhanced electrocatalysis, and efficiently preserving the activity of biomolecules. It would have potential applications in acetylcholine determination in health care biological fields. References [1]. B. Liu, Y. H. Yang, Z. Y. Wu, H. Wang, R. Q. Yu, Sens. Actuators, B, Vol. 104, 2005, pp. 186-192. [2]. F. Gonon, M. Buda, R. Cespuglio, M. Jouvet, J. F. Pujol, Nature, Vol. 286, 1980, pp. 902-903. [3]. S. Burlet, L. Leger, R. Cespuglio, Neuroscience, Vol. 92, 1999, pp. 627-632. [4]. J. R. Cooper, F. E. Bloom, The Biochemical Basis of Neuropharamacology, Oxford University Press, New York, USA, 1991, pp. 190. [5]. B. John, D. Carolina, Medical Biochemistry, Mosby International, London, UK, 2001 pp. 497. [6]. A. Fisher, I. Hanin, C. Lachman (Editors), Alzheimer’s and Parkinson’s Diseases: Strategies for Research and Development, Plenum Press, New York, USA, 1986. [7]. G. Guadra, K. Summer, E. Giacobini, J. Pharmacol. Exp. Ther., Vol. 270, 1994, pp. 277-281. Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 11-18 18 [8]. H. Varoqui, J. D. Erickson, J. Biol. Chem., Vol. 271, 1986, pp. 27229-27232. [9]. D. R. Haubrich, N. Gerber, A. B. Pflueger, M. Zweig, J. Neurochem., Vol. 36, 1981 pp. 1409-1412. [10]. S. R. Lee, M. M. Rahman, M. Ishida, K. Sawada, Trends Anal. Chem., Vol. 28, No. 2, 2009, pp. 196-203. [11]. D. R. Haubrich, N. Gerber, A. B. Pflueger, M. Zweig, J. Neurochem., Vol. 36, 1981, pp. 1409-1412. [12]. Y. Izaki, K. Hori, M. Nomura, Neurosci. Lett., Vol. 258, 1998, pp. 33-38. [13]. M. P. Roisin, J. L. Brassart, G. Charton, Y. Benari, J. Neurosci. Methods, Vol. 37, 1991, pp. 183-188. [14]. T. Yao, S. Handa, Anal. Sci., Vol. 19, 2003, pp. 767-771. [15]. K. F. Martin, C. A. Marsden, Life Sci., Vol. 41, 1947, pp. 865-869. [16]. L. I. Netchiporouk, N. F. Shram, C. Martelet, R. Cespuglio, Anal. Chem., Vol. 68, 1996, pp. 4358-4363. [17]. M. McNally, D. K. Y. Wong, Anal. Chem., Vol. 73, 2001, pp. 4793-4798. [18]. B. J. Venton, K. P. Troyer, R. M. Wightman, Anal. Chem., Vol. 74, 2002, pp. 539-542. [19]. P. Uutela, R. Reinila, R. Kostiainen, Rapid Commun. Mass. Spectrum., Vol. 19, 2005, pp. 2950-2954. [20]. P. E. Potter, J. L. Meek, N. H. Neff, J. Neurochem., Vol. 41, 1983, p. 188. [21]. A. S. Poghossian, Sens. Actuators, B, Vol. 44, 1997, p. 361. [22]. S. R. Lee, Y. T. Lee, K. Sawada, H. Takao, M. Ishida, Biosens. Bioelectron., Vol. 24, 2008, p. 410. [23]. A. Zurn, B. Rabolt, M. Grafe, H. Muller, Fresenius J. Anal. Chem., Vol. 349, 1994, p. 666. [24]. D. G. Pijanowska, W. Torbicz, Sens. Actuators, B, Vol. 44, 1997, p. 370. [25]. A. M. Nyamsi-Hendji, C. Martelet, P. Clechet, Anal. Chim. Acta, Vol. 281, 1993, p. 3. [26]. T. Kullick, U. Bock, J. Schubert, T. Scheper, K. Schugerl, Anal. Chim. Acta, Vol. 300 1994, p. 25. [27]. S. Caras, J. Janata, Anal. Chem., Vol. 52, 1980, p. 1935. [28]. B. H. van der Schoot, P. Bergveld, Biosensors, Vol. 3, 1987, p. 161. [29]. D. Hafeman, J. Parce, H. McConnell, Science, Vol. 240, 1988, p. 1182. [30]. A. P. Soldatkin, D. V. Gorchkov, C. Martlet, N. Jaffrezic-Renault, Mater. Sci. Eng., C, Vol. 5, 1997, p. 35. [31]. K. Sawada, T. Shimada, T. Ohshina, H. Takao, M. Ishida, Sens. Actuators, B, Vol. 98, 2004, p. 69. [32]. C. Y. Tian, N. Q. Jia, R. Wang, Z. Zhang, J. Z. Zhu, G. X. Zhang, Sens. Actuators, B, Vol. 52, 1998, p. 119. [33]. M. M. Rahman, A. Umar, K. Sawada, Adv. Sci. Letters., Vol. 2, 2009, pp. 28–34. [34]. B. N. Barsoum, W. M. Watson, I. M. Mahdi, E. Khalid, J. Electroanal. Chem., Vol. 567, 2004, pp. 277–281. [35]. S. Lin, C. C. Liu, T. C. Choua, Biosensors and Bioelectronics, Vol. 20, 2004, pp. 9–14. [36]. P. C. Pandey, S. Upadhyay, H. C. Pathak, Ida Tiwari, Sens. Actuatos B, Vol. 62, 2000, pp. 109–116. [38]. H. Heli, M. Hajjizadeh, A. Jabbari, A. A. Moosavi-Movahedi, Biosensors and Bioelectronics, Vol. 24, 2009, pp. 2328-2333. ___________________ 2011 Copyright ©, International Frequency Sensor Association (IFSA). All rights reserved. (http://www.sensorsportal.com) http://www.sensorsportal.com/HTML/BioMEMS.htm Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 19-30 19 SSSeeennnsssooorrrsss &&& TTTrrraaannnsssddduuuccceeerrrsss ISSN 1726-5479 © 2011 by IFSA http://www.sensorsportal.com Development of Pressure-Temperature Integrated Multifunction Sensor Using Piezo-Resistive Element 1Palash K. Kundu, 2Gautam Sarkar and 3Chandan Dutta 1Department of Electrical Engineering, Jadavpur University, Kolkata – 700032, India 2Department of Electrical Engineering, Jadavpur University, Kolkata – 700032, India 3Department of Electrical Engineering, Bengal Institute of Technology and Management, Birbhum – 731236, India E-mail: kundupalash2004@yahoo.co.in, sgautam63@yahoo.com, chand.bitm@gmail.com. Received: 4 December 2010 /Accepted: 17 March 2011 /Published: 29 March 2011 Abstracts: A novel attempt was made to develop a multifunction sensor using piezo resistive material for sensing pressure and temperature simultaneously as because it is well known that piezo resistive material has better selectivity to both temperature and pressure or force variables. The advantage of use of piezo resistive material is that it occupies minimum space. The aggregated output, when excited by electrical signal varies with respect to temperature and pressure both. From the output, the temperature and pressure values are extracted with developed model using multiple regression technique and artificial neural network. Copyright © 2011 IFSA. Keywords: Multiple linear regression, Peizo-resistive sensor, Radial basis network. 1. Introduction In recent years, the processing of high level information that is related to many conventional physical or other kinds of quantities or variables has rapidly evolved. A multifunction sensor provides a plurality of physical variables in one sensor module in a minimum occupied space which can interface with and control operation of one or more processor control systems. Thus it can provide also high- level and high dimension information, which is great importance to intelligent systems, in which it is relevant to extract (or to fuse) data from multi-sensor systems in order to acquire more accurate information about the external environment and to decrease the uncertainties that hinder their manipulation. In industrial environment, the multifunction sensors minimize the wiring cost with Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 19-30 20 increased marital complexity and more data processing time. The processing of high level information consists in two aspects, multi sensing strategy and aggression of multi-sensor observations with adaptive algorithms, which are well known as the techniques of sensing fusion. Fusion technique is expected to increase the reliability of measurement (that is fairly immune to noise and to sensor failure). Multifunction sensors are equally important in the field called Bionics, in which human behavior can be studied. Applications for the multi function sensors include automotive test and measurement applications. The topology of the multi-sensor depends on the different aspects as i) multimode sensor in which a given sensor is configured at a time for sensing a particular variable among group of variables, ii) multi variants sensor, in which the given sensor can sense no. of physical variables (more than one) simultaneously at a time. In later case the output of sensor consists aggregation (mixed) of output corresponding to each variable. The topology can be best viewed [4] by the following functional diagram (Figs. 1 and 2). In the 1st type, different sensor materials are chosen for sensing the physical variables (e.g. temperature, pressure, flow etc.). These materials (preferably doped semiconductors) are embedded into a single package. The output of the sub sensors are multiplexed by suitable mode control input signal as shown above. The materials of the sub sensors are fabricated in multiple layers over wafer (base material). In the 2nd type, sensor material, itself has cross sensitivity with respect to multiple physical variables. Thus when these sensors are exposed to change in external physical variables, its output aggregates the partial output corresponding to the change of individual physical variables (e.g. temperature, pressure etc.). External means is adopted to segregate the individual quantity following special algorithm as multiple linear regression method [1], Quantity creditability tactics (QCT) [7]. The evaluation of quantities is subjected to the reconstruction criterion operated on multifunctional measurement equations of the sensors. Almost all the multifunctional sensors are developed to examine no more than two functions. For more than three variables, the reconstruction of quantities must be performed in three dimensional spaces, in which arbitrary choice may be more frequently imposed and may lead to the algorithm of reconstruction invalid. Fig. 1. Pressure and Temperature Sensitive peizo-resistive elements embedded in single package. When C = 0, sensor is selected in pressure sensor mode; when C = 1 sensor is selected in temperature sensor mode. In present study, the experiments are done with fabricated prototype of piezoresistive element for sensing the temperature and pressure simultaneously. The experimental data set related to input output characteristic are utilized for development of model using multiple linear regression method and Radial Basis Neural Network. The prototype sensor is tested with built up model to indicate the temperature and pressure of the source simultaneously. Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 19-30 21 Fig. 2. Pressure and Temperature sensitive elements in different layers. 2. Development of Multi Function Sensor Model 2.1. Multiple Linear Regression In multiple variable sensing, regression analysis may be suited in which there are more than one regression variable. A regression model that contains more than one variable is called a multiple regression model as given by [2]: Y = β0 + β1x1 + β2x2 +  , (1) where Y represents the output and x1, x2 represents the two input variables and  is a random error. This is a multiple linear regression model with two regressor. The term linear means a function of unknown parameters like β0, β1, β2. The above regression model equation describes a three dimensional space of Y, x1, and x2. Fig. 3. shows this plane for regression model. E(Y) = β0 + β1x1 + β2x2 , (2) where we have assumed that the expected value or the error term is zero; that is E (ε) = 0. β0 is the intercept of the plane. β1 and β2 sometimes call partial regression co efficient. Fig. 3. Regression plane. Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 19-30 22 In general, the dependent variable or response Y may be related to k independent regressor variable. Y = β0 + β1x1 + β2x2 + …. + βkxk +  (3) The model is called a multiple linear regression model with k regressor variables. The parameter βj, j = 0, 1… k, are called regression co efficient. This model describes a hyper plane in the k- dimensional space of the regressor variables {xj} as shown in Fig. 3. the parameter βj represents the expected change in response Y per unit change in xj when all remaining regressor xi (i = j) are held constant. Multiple linear regression models are often used as approximation functions. That is the true function relationship between Y and x1, x2 …xk is known but over a certain range of the independent variables the linear regression model is an adequate approximation. An interaction between two variables can be represented by a cross-product term in the model, such as Y = β0 + β1x1 + β2x2 + β12x1x2 +  (4) If we let x3 = x1x2 and β3 = β12, the above equation can be written as Y = β0 + β1x1 + β2x2 + β3x3 +  (5) This is linear regression model. In general, any regression model may be of higher order (i.e. 2nd, 3rd etc.) [2]. As a final example, consider the second order model with interaction Y = β0 + β1x1 + β2x2 + β11x1 2 + β22x2 2 + β12x1x2 +  (6) If we let x3 = x1 2, x4 = x2 2, x5 = x1x2 and β3 = β11, β4 = β22 and β5 = β12, the above equation can be written as a multiple linear regression model as follows: Y = β0 + β1x1 + β2x2 + β3x3 + β4x4 + β5x5 +  (7) The method of least square may be used to estimate the regression co efficient in the multiple regression models, in equation 3. Suppose n > k observation are available, and let xij denote the ith observation or level of variable xj, the observation are: (xi1, xi2… xik, yi), I =1, 2….n and n> k. It is customary to present the data for multiple regressions in a table. Each observation (xi1, xi2… xik, yi), satisfies the model equation (3) or,    k j jijjikkiii XXXXy 1 0122110 .....  for i = 1, 2…n (8) The least square function is 2 2 0 1 1 1 n n k i j i i j L yi ij                  x (9) We want to minimize L with respect to β0, β1… βk. The least square estimation of β0, β1… must satisfy Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 19-30 23 0)(2 1 0 1,,0 ,.......210      k j ijj n i i xy L k      and kjxxy L k j ijijj n i i j k ,....2,1,0)(2 1 0 1 ,, ,.......210           (10) Simplifying the above equation, the least square normal equations associated with (k+1) nos of unknown regression coefficients (i.e. s' ) are obtained. The solution to the normal equations will be least square estimators of the regression coefficients. In fitting a multiple regression model, it is much more convenient to express the mathematical operations using matrix notation. Suppose that there are k regressor variables and n observations, (xi1, xi2, p, xik, yi), i = 1, 2, p, n and that the model relating the regressor to the response is Yi = β0 + β1xi1 + β2xi2 + … + βkxik + 1 i =1, 2… n This model is a system of n equations that can be expressed in matrix notation as: Y = Xβ +  , where, Y = 1 2 3 y y y              , X = 11 12 1 21 22 2 1 2 nk 1 x x x 1 x x x 1 x x x k k n n                    , β = 0 1 k                 and 0 1 k             Now, the vector of least squares estimators, β, as to be found, that minimizes )()( '' 1 2  XYXYL n i i    (11) The least squares estimator   is the solution for  in the equations: 0    L : Hence the solution of the above equation, YXXX ''   (12) The above equations are called the least squares normal equations in matrix form. Solving these normal equations, the least squares estimate of  is obtained as Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 19-30 24 YXXX .).( '1'  (13) Note that there are p = k +1 normal equations in p = k+ 1 unknowns (the values of Furthermore, the matrix X’X is always nonsingular, as was assumed above, so the methods described in textbooks on determinants and matrices for inverting these matrices can be used to find. In practice, multiple regression calculations are almost always performed using a computer. It is easy to see that the matrix form of the normal equations is identical to the scalar form. Writing out Equation (10) in details we obtain 1 2 1 1 1 2 1 1 1 2 1 1 1 1 1 2 1 2 1 1 1 1 n x x x x x x x x x x x x x x x n n n i i ik i i i n n n n i i i i i ik i i i i n n n n ik i ik i ik ik i i i i                                                   0 1 k                           = 1 1 1 1 n i i n i i i n ik i i y x y x y                             If the indicated matrix multiplication is performed, the scalar form of the normal equations (i.e., equation (13)) will result. In this form 'X X is a (pp) symmetric matrix and 'X y is a (p1) column vector. Note the special structure of 'X X the matrix. The diagonal elements 'X X are the sums of squares of the elements in the columns of X, and the off-diagonal elements are the sums of cross-products of the elements in the columns of X. Furthermore, note that the elements of are the sums of cross-products of the columns of X. and the observations { 'e y y  }. The fitted regression model is: 0 1 ' k i j j j y x     , i= 1,2,…,n In matrix notation, the fitted model is: .' XY  The difference between the observation yi and the fitted value yi’ is a residual, say, ' i i ie y y  . The (n1) vector of residuals is denoted by 'e y y  2.2. Radial basis Neural Network Radial Basis functions emerged as a variant of artificial neural network in late 80’s. [2]. However, their roots are entrenched in much older pattern recognition techniques as for examples potential functions, clustering, functional approximations, spline interpolation and mixture models. In topology, the Radial basis functions are embedded into three layers feed forward networks. In between the inputs layers and output layers there is a layer of processing units called hidden units. Each of them implements a radial basis function. The output, , of the network is thus , Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 19-30 25 where N is the number of neurons in the hidden layer, is the center vector for neuron i, and ai are the weights of the linear output neuron. In the basic form all inputs are connected to each hidden neuron. The norm is typically taken to be the Euclidean distance and the basis function is taken to be Gaussian The Gaussian basis functions are local in the sense that i.e. changing parameters of one neuron has only a small effect for input values that are far away from the center of that neuron. RBF networks are universal approximators on a compact subset of . This means that a RBF network with enough hidden neurons can approximate any continuous function with arbitrary precision. The weights ai, , and β are determined in a manner that optimizes the fit between and the data. The topology of Radial Basis Network is shown in Fig. 4. Fig. 4. Topology of Radial Basis Network. The || dist || box in this figure accepts the input vector x and the input weight matrix IW1,1, and produces a vector having S1 elements. The elements are the distances between the input vector and vectors iIW 1,1 formed from the rows of the input weight matrix. The bias vector b1 and the output of || dist || are combined with element-by-element multiplication. 3. Experimental Set Up for Multi Sensor Calibration The piezo-resistive sensor has been configured as multi function sensor for sensing temperature and pressure variables. It is selective to both temperature and pressure variation. The prototype sensor is placed into a air tight environmental chamber and is excited by the flow of constant current source. The pressure is also varied by filling compressed air maintained desired pressure. For each set pressure, the temperature of inside space is varied by temperature control facility. For change in Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 19-30 26 temperature by 1C and pressure by 1 psi, the output (e.m.f.) of the sensor is recorded by computer based data acquisition system and control. The observations are tabulated in Table 1(a). Fig. 5. Experimental set-up for sensor calibration. Table 1(a). Observed output (emf) from piezo –resistive sensor system. Temperature (0 C ) Pressure (p.s.i.) 10 20 30 40 50 60 70 15 -0.8627 -0.8100 -0.7451 -0.7075 -0.6722 -0.4700 -0.6546 16 -0.2255 -0.2360 -0.1760 -0.1440 -0.1147 -0.0932 -0.1050 17 0.3501 0.3370 0.3915 0.4182 0.4425 0.4604 0.4450 18 0.9211 0.9117 0.9500 0.9796 0.9999 1.0131 0.9947 19 1.4963 1.4855 1.5225 1.5411 1.5570 1.5662 1.5430 20 2.0718 2.0574 2.0925 2.1035 2.1136 2.1190 2.0927 21 2.6469 2.6298 2.6584 2.6644 2.6700 2.6718 2.6420 22 3.2213 3.1993 3.2233 3.2246 3.2252 3.2225 3.1890 23 3.7966 3.7691 3.7877 3.7850 3.7504 3.7743 3.7376 24 4.3686 4.3399 4.3521 4.3447 4.3366 4.3251 4.2847 25 4.9411 4.9092 4.9164 4.9045 4.8919 4.8763 4.8322 26 5.5135 5.4783 5.4801 5.4638 5.4402 5.4273 5.3799 27 6.0859 6.0466 6.0440 6.0226 6.0013 5.9786 5.9271 28 6.6588 6.6141 6.6065 6.5810 6.5556 6.5281 6.4741 29 7.2286 7.1818 7.1691 7.1397 7.1093 7.0781 7.0205 30 7.7995 7.7486 7.7313 7.6976 7.6626 7.6288 7.5872 4. Testing and Simulation of Prototype Sensor A complete prototype pressure–temperature transducer is embedded in a single package containing piezo-resistive element placed in one arm of four arms Wheatstone bridge circuit and is excited by 5 V D.C. power supply under constant current condition. The output (i.e. error voltage) from the bridge is adjusted by gain amplifier within the full scale range (0 to 5 V) and is converted by in-built A/D convertor of PIC microcontroller (i.e. 16F877) driven by 16 MHz external XTLA as clock oscillator. The emf, (y) are measured by varying temperature (x2) and pressure (x1) at intervals of 1C and 1psi within full scale range as 0-70 C and 15-30 psi. by experimentation. The observations are tabulated (Table 1(a)). Using multiple linear regression model of n-th order the coefficient constants are found out. The 1st order multiple linear regression model is given by:   2211 xxy O Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 19-30 27 The 2nd order model is given by:   2514213 2 22 2 11 xxxxxxy O , where y is the e.m.f, vector x1 is the temperature vector, and x2 is the pressure vector. According observations as shown in Table 1(a), applying matrix approach, the fitted regression model is determined from least square estimators. The least square estimates are found from equation as: 1( ' ) 'X X X y  , where X=[x1, x2] 0 1 2             = -9.3257 0.5701 -0.0001          Therefore, the fitted regression model with the regression co-efficient place is 2 2' -9.3257 0.5701 (-0.0001)y x x   , where y’ is the computed e.m.f. The residual corresponding to the first observation is ' 1 1 1e y y  Extraction of temperature and pressure variables as output from emf as input is done with the help of 1st order multiple linear regression model and ANN model as shown in the following block diagram (Fig. 6) The ANN-I model is developed by training with emf (input variable) and temperature/pressure ratio (target variable). (Table 1(b)). After successful training, it is tested with testing data set containing unknown emf (input) and the output of the ANN-I produced predicted ratio ( 1 2 x x r  ), which is passed to 1st order multiple linear regression model, already developed. The temperature and pressure variables are solved using set of equations as given below: 1 2 x x r  , or 12 rxx  , also r x x 2 1  , Substituting the value of 1x (i.e. r x x 2 1  ) into regression model equations, we obtain, 22 2 1 x r x y O   , 2 1 2       r y x O Similarly substituting the value of 2x (i.e. x2 = rx1), we obtain, 1211 xrxy O   21 1   r y x O    Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 19-30 28 Fig. 6. PIC based DAS for piezo resistive sensor interface with computer. Table 1(b). Computed ratio (r) against different values of pressure and temperature. Temperature ( 0C ) Pressure (p.s.i.) 10 20 30 40 50 60 70 15 0.666 1.333 2.000 2.666 3.333 4.000 4.666 16 0.625 1.250 1.875 2.500 3.125 3.750 4.375 17 0.588 1.176 1.764 2.352 2.940 3.528 4.116 18 0.555 1.111 1.665 2.222 2.775 3.330 3.885 19 0.526 1.052 1.578 2.104 1.578 3.156 3.682 20 0.500 1.000 1.5 2.000 2.500 3.000 3.500 21 0.476 0.952 1.428 1.904 2.380 2.856 3.332 22 0.454 0.412 1.362 0.824 2.27 2.724 3.178 23 0.434 0.868 1.302 1.736 2.17 2.604 3.038 24 0.416 0.832 1.248 1.664 2.08 2.496 2.912 25 0.400 0.800 1.200 1.600 2.00 2.400 2.800 26 0.384 0.768 1.152 1.536 1.920 2.304 2.688 27 0.370 0.740 1.110 1.480 1.850 2.220 2.590 28 0.357 0.714 1.071 1.428 1.785 2.142 2.499 29 0.344 0.688 1.032 1.376 1.720 2.064 2.408 30 0.333 0.666 0.999 1.332 1.665 1.998 2.331 Sensors & Transducers Journal, Vol. 126, Issue 3, March 2011, pp. 19-30 29 Now, since the ANN-I model has been trained with computed values as obtained using 1st order regression model equation, the predicted ratio may include error. This is mainly because of the non linear cross sensitivity for two variables of the multi function sensor. This error has to be nullified by 2nd ANN model (ANN-II). The ANN-II has been trained using actual temperature and pressure variables and corresponding error in them as obtained from error analysis data set. During testing and simulation, the computed pressure and temperature are passed into input of the ANN-II model and are corrected after summing predicted errors in pressure and temperature obtained from the output of the ANN-II model as shown in the following block diagram. Here all ANN model are structured as Radial Basis Neural Network (RBNN) [2]. Fig. 7. Model for extraction of pressure and temperature values from sensor output.