Review article: Making environmental sensors on plastic foil

With the emergence of the printed electronics industry, the development of sensing technologies on non conventional substrates such as plastic foils is on-going. In this article, we review the work performed and the trends in the development of environmental sensors on plastic and flexible…

With the emergence of the printed electronics industry, the development of sensing technologies on non conventional substrates such as plastic foils is on-going. In this article, we review the work performed and the trends in the development of environmental sensors on plastic and flexible foils. Our main focus is on the integration of temperature, humidity, and gas sensors on plastic substrates targeting low-power operation for wireless applications. Some perspectives in this dynamic field are also provided showing the potential for the realization of several types of transducers on substrates of different natures and their combination with other components to realize smart systems.

Danick Brianda*, Alexandru Opreab, Jérôme Courbata, and Nicolae Bârsanb
a Ecole Polytechnique Fédérale de Lausanne (EPFL), Institute of Microengineering, Sensors, Actuators and Microsystems Laboratory, Rue Jaquet-Droz 1, P.O. Box 526, CH-2002, Neuchâtel, Switzerland
b Institute of Physical and Theoretical Chemistry, University of Tübingen, Auf der Morgenstelle 15, D-72076 Tübingen, BW, Germany
* E-mail:

The driving forces for organic and printed electronics are the display and lightning, solar cell, battery and electronics (e.g., RFID) industries1. The complete technology chain is being established in the fields of materials preparation, processing and characterization equipment, and production. Over the last decade, there has been a significant increase in the efforts dedicated to the development and implementation of electronic components on flexible and
stretchable substrates for other types of application, such as sensing, and notable results have been obtained by different research groups2-8. This technology could result in sensors being introduced to new settings, by significantly reducing their production cost and by adding new functionalities.

The monitoring of environmental parameters in a distributed manner is of significant interest in different fields, for comfort, environmental, health, safety, and security purposes. A lot of work is underway on the development of smart sensors and wireless sensor networks based on silicon technology, targeting different types of application. Printed electronics are becoming a more and more mature technology every day, and new kinds of product are expected in the near future. Besides a strong potential for cost-effective production based on additive processes with a reduced infrastructure, the benefits of printing devices on plastic foil include their potential to be light weight, foldable/rollable, transparent, thin and conformal, wearable, and produced on a large scale, depending on the processing technology involved.

In this review, we report on the recent advances achieved in the development of individual sensors and multi-sensor platforms on plastic foil for environmental monitoring, with a special focus on temperature, humidity, and gases. We will begin by reviewing the exciting work that has been performed in this field in recent years. Secondly, we will introduce some of the work we carried out on this topic, with the integration of different environmental sensors on a single plastic platform. We have integrated different sensing principles on a polyimide foil, such as capacitive and resistive read-outs for the detection of several types of environmental parameters including temperature, humidity, reducing and oxidizing gases, and volatile organic compounds (VOCs). These sensors on plastic foils are required to realize intelligent RFID tags for environmental monitoring9. Such devices may eventually find application in  wearable systems, smart buildings, and in the logistics of perishable products.


Gas sensors on plastic foil

Different transducing principles have been developed for atmospheric gas sensing10. These principles include the resistive principle, mainly based on metal-oxide and polymeric (chemiresistors) gas sensitive films; the capacitive principle, involving a change in the dielectric constants and/or a swelling of the sensing film; the field-effect principle, based on a change of work function and semiconductor surface potential; the colorimetric principle, in which the optical absorption spectrum is modified by the gaseous analyte; and the resonating principle, in which an addition of mass modifies the resonant frequency of the resonator. Most of the gas responses of these devices significantly depend on the temperature and humidity content of the surrounding environment. Moreover, the gas sensors based on these transducing principles suffer from a lack of selectivity. It is therefore constructive to use an array of gas sensors alongside temperature and humidity sensors to obtain valuable information on the composition of the surrounding atmosphere. Considerable efforts have been dedicated to the miniaturization of these transducers based on silicon technology. A nice example of their integration into arrays with the electronics interfaced on a single silicon chip has been produced at the ETHZ in Switzerland11.

There are a limited number of publications on gas sensors on plastic/flexible foils, but this number is growing. Interest in this area began with the development of organic electronic transistors and the study of issues surrounding their sensitivity to humidity and different gases; this has necessitated investigations on humidity and gas impermeable encapsulation barriers, as well as air stable organic semiconductors. Some groups saw an opportunity to exploit this drawback to make gas and humidity sensitive devices. They have worked on organic thin film transistor (OTFT) based gas sensors: on single devices and arrays, on silicon and plastic substrates, for sensing volatile organic compounds (VOCs)2,7,13-15. One important achievement has been reported by Torsi et al., in the form of a novel chiral bilayer organic thin-film transistor gas sensor, comprising an outermost layer with built-in enantioselective properties that exhibits a field- ffect amplified sensitivity that enables differential detection of optical isomers in the tens-of-parts-per-million concentration range7. However, the gas sensors based on organic transistors require further development to achieve the required sensing performance and reliability needed for commercialization, as discussed by the group of V. Subramanian at University of California, Berkeley in reference16.

The development of other types of gas sensor on plastic-flexible substrates has only begun very recently. Most of the samples are made on polyethylene- erephthalate (PET), polyethylene naphthalate (PEN), and polyimide (PI), and some on parylene substrates. Conventional and printed hybrid processes and organic and inorganic hybrid materials are generally used with the aim of producing fully print compatible devices, as conceptually illustrated in Fig. 1. A major part of the published work consists of single flexible humidity sensors, sometimes combined with a temperature sensor on the same platform17-24. Capacitive and resistive transducers are commonly used as sensor architectures. Volatile organic compounds ammonia and hydrogen sulphide sensors have also been reported, some of which use printing technologies to deposit gas-sensitive conducting polymers and silver electrodes25-29. Articles have also been published on NOx detection in the sub and low ppm range based on ink-jet printed inorganic or organic polymeric materials, amorphous OTFT operating at room temperature, and resistive PEDOT:PSS; however, the gas sensing performance has been poor30,31. The colorimetric detection of gases on foil has been also demonstrated with the use of gas sensitive dyes combined with an optical waveguide on plastic, with the detection of concentrations of CO2 below one percent in nitrogen, and the sub-ppm detection of NH3 in air32,33. Looking at the different sensing principles described above, only resonating type gas sensors have not yet been fabricated directly on plastic foil. A hybrid approach in which a surface acoustic wave (SAW) chip has been transferred onto a plastic substrate has been reported, but was applied to light sensing in that communication34. Another interesting and new approach is the coating of passive (no power source on board) conventional RFID tags with chemically sensitive films to form a chemical sensor35. The detection of several vapors of industrial, health, law enforcement, and security interest (ethanol, methanol, acetonitrile, and water vapors) was demonstrated with a single 13.56 MHz RFID tag coated with a solid polymer electrolyte sensing film. For multicomponent detection and quantification using a single RFID sensor, multiple parameters from the measured real and imaginary portions of the complex impedance are calculated. Finally, some groups have started to look at the implementation of gas sensitive nanomaterials on plastic foil with the transfer or self-assembly of nanowires, nanotubes, and nanoparticles on flexible substrates36-42.

Regarding the integration of temperature sensors on plastic-flexible foil, the conventional platinum resistance temperature detector (RTD) has been realized on a flexible polyimide substrate, for operation up to 400 °C, and resistors made of TaSiN have been shown to exhibit high temperature coefficient of resistance (TCR) values43,44. Some approaches based on thermo-sensitive polymers have also been evaluated45,46, based on graphite or metallic powders in a PDMS matrix, which suffered from non-linearity and were limited to 100 °C. Another approach that has been reported is the screen printing of a polymeric thermo-sensitive material on Kapton for textronic applications, e.g., measurement of the temperature of the human body47.

In the process of considering the next generation of smart gas sensors (besides silicon based technologies) EPFL-IMT SAMLAB has launched GASID (GAS IDentification, in reference to RFID) to look at the potential integration of micro gas sensors on plastic foil. In collaboration with the Institute of Physical Chemistry at the University of Tübingen and the Fraunhofer IPM in Germany, this initiative has led to the proof of concept for capacitive differential VOCs/humidity sensors, semi-conductor metal-oxide gas sensors, and colorimetric gas sensors on plastic foil (PET, PEN, and PI)33,48,49. Surprisingly enough, EPFL-IMT SAMLAB and the University of Tuebingen were able to demonstrate the continuous operation of metal oxide gas sensors made on polyimide hotplates for several months48. They have also shown, using a simple sensor architecture and making use of the plastic substrate as humidity sensing element, that volatile organic compounds and humidity can be measured simultaneously using two capacitive sensors in differential measurement mode49. EPFL-IMT SAMLAB and the University of Tuebingen have produced a multi-parameter sensing platform (for VOCs, temperature, humidity, reducing and oxidizing gases) on plastic foil, based on standard clean room processes48. These devices have great potential but their manufacture has to be rethought since plastic substrates are not welcome in conventional microelectronics foundries. Production at the lowest possible cost is vital in order to open up the market, which is currently inaccessible thanks to silicon sensor technologies. To reach the cost targets, one needs to develop heterogeneous materials, processes, and integration methods to enable the development of multi gas sensor platforms.

The work presented in the next section most likely represents the most advanced assessment of the performances of multi-parametric sensing devices on plastic foils. Meanwhile, the reliability aspects and the evaluation of their flexibility have yet to be fully addressed, with studies on reliability only recently being released in the field of flexible electronics50,51.


Multi-parametric sensing platforms

The simultaneous detection and quantification of physical, chemical, and biological information from the ambient with mobile/autonomous/ remote sensing systems is easier accomplished when using complex platforms that integrate several dissimilar sensors; ideally all those required by a specific application. The first step towards flexible multi-parametric sensing platforms should be, and actually was, the development of different kinds of sensors on plastic substrates. Successfully attempts have been already made for humidity23, reducing or oxidizing gases48,52,53, and volatile organic compounds (VOCs)49. In reference52 the integration of metal oxide (MOX) based gas sensors for reducing and oxidizing gases is described (see Fig. 2). The results obtained were promising, as the response of the sensor on polyimide (PI) foil was, analyte depending, between 40 % and 100 % from that of a reference sensor produced on a silicon nitride hotplate using the same deposition method (drop coating)54. Principally intended to demonstrate the sensor concept viability, the hotplates on flexible PI arrays have been coated with only one sensing material: SnO2.

A diversification of the sensor types on one PI platform is reported in references48,53. In reference48 two different types of MOX (SnO2 and WO3) have been deposited on PI platforms containing several transducing areas. Hydrophobic Teflon based filtering layers (see Fig. 3) have been employed to increase the selectivity. The foil level packaging of the chemical gas sensors is described in detail in reference55. In order to reduce the readout power, the same contribution proposes the direct sensor readout on the sensor system microcontroller using the time constant of an RC circuit that includes the sensor resistor. The heating power could be also reduced from 13.7 mW to 340 μW for one hotplate through pulse operation53.

In the work presented in reference49, a Pt-resistance thermometer and two additional capacitive interdigital structures have been patterned in the same processing step (see Fig. 4). The number of capacitors is not technologically limited, as shown later on, but two are enough to underline the principle of operation.

Coated with suitable polymers the capacitors can independently detect different VOCs, provided the substrate sensitivity towards the analytes is significantly lower than that of the sensing layers. Often this requirement is not satisfied by plain electrotechnical-grade foils and in these cases the full platform potential is coming into play. In order to eliminate the effects of the unwished residual sensitivity of the substrates to gases, only one capacitor is covered with functionalized polymers, the other remaining uncovered and delivering a capacitance reference. Using a differential readout, which can be directly implemented at hardware level through a differential capacitance to digital converter, the response of the substrate foil is canceled out from the useful signal. Thus the “smartness” of the sensor system relies directly on the sensors themselves and not on the software driving the system microcontroller.

The operation of the platform can be easily understood by analyzing the sensor responses to controlled changes of the ambient atmosphere composition, by using a measuring chamber with a gas mixing system. As depicted in Fig. 5, the standard evaluation and calibration procedure is based on several independent exposures towards test VOCs/gases (n-hexane, n-propanol, ethanol, toluene, ammonia and humidity) diluted in dry synthetic air (80 % N2 + 20 % O2 – carrier gas) or in synthetic air with a certain humidity content (humidity background). The test gas concentrations usually start from the time weighted average (TWA) values. Between exposure sequences recovery times are allowed, when only the carrier gas with background humidity is purged through the measuring chamber. Fig. 6a displays the raw capacitance signals from an individual capacitive sensor and reference capacitor. One has to first remark on the huge responses to humidity (caused by the high dipolar momentum of the water molecules), the large time constants associated with the humidity changes, and the apparent lack of response for toluene and ammonia. However, the difference between the sensor capacitance and the reference one (actually the output of the platform in the differential operation mode) results in a quite unexpected picture  (see Fig. 6b). The platforms are sensitive enough to all analytes to allow the extraction of the calibration curves over roughly three orders of magnitude for the concentration of the target gases49. They are reversible and relatively fast (response and recovery times on the order of minutes) but not very selective. The cross sensitivity to humidity, obvious in Fig. 6, drastically reduces the performance of single platforms.

By using the same concept, capacitive sensor arrays have been realized56, which in conjunction with suitable recognition software provide reasonable predictions concerning the composition of the gaseous/VOCs mixtures (see Fig. 7). For demonstration proposes the Unscrambler® program has been used, but dedicated software would be required for practical applications. The success of linear algebraic methods, based on linear sensor responses, is often compromised by the cross sensitivity of the devices. By losing the simplicity and increasing the cost, it is possible to foresee the implementation of more complicated mathematical approaches.

In some cases, the “parasitic” capacitive contributions of the substrates (dashed olive curve in Fig. 6a) can play a positive role, acting as a humidity sensor in applications where the response and recovery times are not critical49.

A more complex sensor integration approach48 brings together temperature, MOX and capacitive gas/humidity/VOCs sensors on the same PI substrate by combining the sensors addressed above. In order to miniaturize the multi-sensor platform the capacitive and temperature sensors have been scaled down (see Fig. 8).

The signals of all gas sensors acquired during a demonstrative exposure to NO2 (oxidizing gas), ethanol (reducing gas) and humidity, are depicted in Fig. 9. Panel (a) shows the sensor behavior at low ethanol concentration while panel (b) refers to a higher ethanol concentration. In the high ethanol concentration range the capacitive response of the platform to ethanol is visible, in addition to that for humidity, which is always present. The middle graph of each panel reassesses the extraction procedure of the capacitive response in the differential operation regime indicating, at the same time, the possibility to use the reference capacitor as a humidity sensor. The lower graphs dedicated to the MOX devices point out the respectable sensor functionality, especially with 50 % humidity background that is the normal for environmental applications. One has to remark on the increased sensitivity of the nanogranular WO3 sensing element for NO2 (~70 ppm-1 @ 1 ppm NO2) and its significantly reduced sensitivity for the other analyte, ethanol, resulting in a fair selectivity. Using the TWA values for NO2 and ethanol (5 ppm and 500 ppm respectively) as reference concentrations, one obtains a response (R) ratio (quantifying the selectivity) of: [EQUATION] .

The overall merit figure of the SnO2 device (also a nanogranular material) as ethanol sensor is not as good (ethanol sensitivity of ~1 ppm-1 @ 20 ppm ethanol and a response ratio of ~3 at TWA concentrations). The dissimilarities between the two MOX sensor performances are related to the sensing material and analyte characteristics, but also to the fact that the manufacturing technology was not optimized; the communication proposing only a sensor concept and proving its feasibility. The influence of the humidity on the MOX sensors is reduced if the variations occur in the middle and upper humidity ranges (30 % to 90 % relative humidity) due to the rather high operation temperature (280 °C). However, alternation of the exposure sequences with and without humidity backgrounds impinges on the sensing layer surface properties (through the superficial concentration and type of the OH groups) and results in some baseline (sensor signal in the absence of the main analyte) drifts that can be observed in Fig. 9. In spite of these drifts, the response reproducibility is good, with mid and long term stability for both sensors types (MOX and polymer based). A new contribution containing statistical data over several months of operation has been prepared and submitted for publication elsewhere. The power consumption of the sensor heater in continuous operation mode was about 18 mW but intermittent or pulse operation are also possible as mentioned above. Through the examples given, “flexible” and “on foil” environmental sensor systems have been revealed as feasible, and are pushing research and development interest/activities/efforts towards cheaper and large scale manufacturing technologies. This type of approach will require new material structures and morphologies, compatible with the new production tools and conditions, and require continuous feedback from materials science.



It is foreseeable that the fabrication of physical and chemical sensors on plastic foils will evolve towards all printable technological solutions. For that to happen, one needs the formulation of the appropriate inks, especially for the chemically sensitive materials. As already seen in this review, polyimide will only be used for applications with specific requirements regarding temperature and the robustness of the substrate: most devices will be produced on PET and PEN substrates. Regarding the type of sensing devices addressed here, the ink-jet printing of electrodes and sensing layers, organic and inorganic, is underway in different groups. The use of plastic substrates is also compatible with the low temperatures required for the preparation of nanostructures and their functionalization with chemical and biological agents.

The emerging industry of large area manufacturing and organic & printed electronics is bringing about new opportunities for the realization of sensors on unconventional substrates that could lead to new applications in the near future. On one hand, printed electronics are being considered as a production means for very low-cost RFID tags. On the other hand, there is a need for a variety of cost-effective sensors that could be manufactured directly on RFID labels to make them smarter; not only temperature, humidity and gas sensors, but also accelerometers (vibrations, shock), light and pressure sensors, to name a few.

The heterogeneous integration of components into smart sensing systems will be a key aspect in the future development of these sensing devices. The development of memories, power sources, and communication components (e.g., antennas) on plastic foil and their integration is on-going. A System in Foil approach could allow the integration of all these components on a unique foil or on different foils “laminated” together, but many issues remain regarding the production yield and reliability, especially under mechanical deformation, of the individual components and the systems.

These smart labels are expected to have an impact in the logistic sector with the monitoring of goods during their transport. In a longer term perspective, cost-effective smart sensing labels on plastic can be envisioned as a key enabling technology in the deployment of the “Internet of Things”. We can imagine that these devices could not only be made on plastic but also on other types of flexible substrates such as paper, thin metal sheets, textiles, biodegradable materials.



Some examples of sensors made on plastic foil have been reviewed in this paper with a specific focus on temperature, humidity, and gas sensors. We have also introduced our work on a multi-sensor platform on flexible polyimide foil that has been developed for the environmental monitoring of different parameters. The characteristics of these platforms are of high interest for the realization of ultra-low power devices that could be processed at low-cost using printing processes. Our next steps are focused on the fabrication of devices using printed processes and their direct integration onto flexible plastic RFID smart labels based on a System in Foil approach, using a combination of organic/printed and inorganic/silicon based components. At EPFL-IMT SAMLAB, we are also looking at the integration of other physical sensors (accelerometers, pressure sensors, resonators) to widen the applications of these smart labels.

There are surely a wider range of sensors and potential applications that may make use of production on plastic substrates. Besides displays, lightning panels, photovoltaic cells, batteries, and OTFTs circuitries, the coming years will lead to a generation of lightweight, flexible, conformable and even transparent sensing devices manufactured on compliant substrates of different natures. The direct printing of devices onto a product is even foreseeable for specific applications.



We are grateful to the Marie-Curie Initial Training Network program under the FlexSmell project (FP7 – Grant ITN no.2 38454) and the GOSPEL Network of Excellence on Artificial Olfaction and Gas Sensing Technologies (FP6 – rant IST no.507610) for the partial funding of the work performed by the authors. 



1. OEA roadmap for organic and printed electronics, 3rd Ed., White paper from the Organic Electronic Association (2009) 78.
2. Zhu, Z. T., Appl Phys Lett (2002) 81(24), 4643.
3. Lacour, S. P., et al., Appl Phys Lett (2003) 82(15), 2404.
4. Khang D. -Y., et al., Science( 2006) 311, 208.
5. Sekitani T. et al., Nat Mater (June 2007) 6, 413.
6. Berggren, M. et al., Adv Mater (2007) 19(20), 3201.
7. Torsi, L., et al., Nat Mater (2008) 7(5), 412.
8. Liu X. Y., et al., Proceedings of MEMS 2011, Cancun, Mexico, January 23-27 (2011) pp. 133.
9. Gadh, R., et al. Eds, RFID : A unique radio innovation for the 21st Century, Proc IEEE (2010) 98(9), 1150.
10. Göpel, W., et al. Eds, Sensors: A comprehensive survey. Chemical and biochemical sensors Part I, VCH Verlagsgesellschaft mbH, Germany (1991) 716 .
11. Hierlemann, A., Integrated Chemical Sensors Systems in CMOS Technology, Springer ( 2005) IX, 225.
12. Mabeck, J. T., et al., Anal Bioanal Chem (2006) 384(2), 343.
13. Liao, F., et al, Sensor Actuat B (2005) 107, 849.
14. Jeong, Y. T., et al., Appl Phys Lett (2008) 93(13), 133304.
15. Subramanian, V., et al., IEEE Transactions on components and packaging technologies (2005) 28(4), 742.
16. Lee, J. B., et al., MRS Symposium Proceedings (2005) 871, 6.
17. Ki, Y. S., Sensor Actuat B (2006) 114, 410.
18. Huanget, A., et al., Sensor Actuat B (2006) 116, 2.
19. Miyoshi, Y., et al., Sensor Actuat B (2009) 142, 28.
20. Arena, A., et al., Microelectr J (2009) 40, 887.
21. Su, P. -G., Sensor Actuat B (2009) 137, 496.
22. Lee, C. -Y., et al., Sensor Actuat A (2008) 147, 173.
23. Oprea, A., et al., Sensor Actuat B (2008) 132, 404.
24. Zampetti, E., Sensor Actuat B (2009) 143, 302.
25. Cho, N. -B., et al., Sensor Actuat B, (2008) 130, 594.
26. Crowley, K., et al., Talanta (2008) 77(2), 710.
27. Su, P. -G., et al., Talanta (2009) 80, 763.
28. Weng, B., Analyst (2010) 135, 2779.
29. Mabrook, M. F., et al., IEEE Sens J (2006), 6(6) 1435.
30. Lin, C. -Y., et al., Sensor Actuat B (2009) 140, 402.
31. Marinelli, F., et al., Sensor Actuat B (2009) 140, 445.
32. Mayr, T., et al., Analyst (2009) 134, 1544
33. Courbat, J., et al., in Proc. of the IEEE Conference on MEMS – MEMS 2010, Wong, M., and Suzuki, Y., (eds.), IEEE, Hong-Kong (2010) 883.
34. Ho, H., et al., in Proc. of the Transducers 2009 conference – 15th International Conference on Solid-State Sensors, Actuators and Microsystems, Najafi, K., et al., (eds.), IEEE, Denver, USA (2009) pp. 1853-1856.
35. Potyrailo, R. A., et al., Anal Chem (2007) 79, 45.
36. Parikh, K., et al., Sensor Actuat B (2006)113, 55
37. McAlpine, M. C., et al., Nat Mater (2007) 6, 379.
38. Pi-Guey, S., et al., Sensor Actuat B (2009) 139, 488.
39. Ahn, H., et al., Electrochem Solid-State Lett (2010) 13(11) J125.
40. Jeong, H. Y., et al., Appl Phys Lett (2010) 96, 213105.
41. Wang, Y., et al., Sensor Actuat B (2010) 150, 708.
42. Wang, L., et al., J Mater Chem (2010) 20, 907.
43. Moser, Y., et al., J Microelectromech S (2007) 16(6), 1349.
44. Chung, C. K., et al., Sensor Actuat A (2009) 156, 323.
45. Shih, W. -P., et al., Sensors (2010) 10, 3597.
46. Chuang, H. -S., et al., J Micromech Microeng (2009) 19, 45010.
47. Bielska, S., et al., Mater Sci Eng B (2009) 165, 50.
48. Courbat, J., et al., Procedia Chem (2009) 1, 597.
49. Oprea, A., et al., Sensor Actuat B (2009) 140, 227.
50. Merilampi, S., et al., Microelectron Reliab (2009) 49, 782.
51. van der Sluis, O., Microelectron Reliab (2009) 49, 853.
52. Briand, D., et al., Sensor Actuat B (2008) 130, 430.
53. Courbat, J., et al., in Proc. of the Transducers 2009 conference – 15th International Conference on Solid-State Sensors, Actuators and Microsystems, Najafi, K., et al., (eds.), IEEE, Denver, USA (2009) 584.
54. Briand, D., et al., Sensor Actuat B (2000) 68, 223.
55. Courbat, J., et al., J Micromech Microeng (2010) 20, 055026.
56. Oprea, A., et al., in Proc. of the Eurosensors XXII conference, Gerlach, G., et al., (eds.), VDI, Dresden, Germany (2008) 1431.