Scholarly article on topic 'Recent advances in printed sensors on foil'

Recent advances in printed sensors on foil Academic research paper on "Materials engineering"

CC BY-NC-ND
0
0
Share paper
Academic journal
Materials Today
OECD Field of science
Keywords
{}

Abstract of research paper on Materials engineering, author of scientific article — Giorgio Mattana, Danick Briand

In this review paper, we summarise the status and trends in the research and development of printed sensors on foil substrates. Our focus includes sensor technologies that have some of their elements printed with a special interest for fully printed structures. The paper reviews the two large physical and chemical sensor families addressing different transduction principles. The paper concludes with a short notice on status and perspectives in the field with some words on the commercial maturity and trends of printed sensors on foil.

Academic research paper on topic "Recent advances in printed sensors on foil"

Materials Today • Volume 00, Number 00• September 2015

RESEARCH

.„JSitrfxmmg.

ELSEVIER

Recent advances in printed sensors on foil

Giorgio Mattana1 and Danick Briand2 *

1 ELORGA - IMS, Bordeaux, France

2 EPFL-IMT SAMLAB, Neuchatel, Switzerland

In this review paper, we summarise the status and trends in the research and development of printed sensors on foil substrates. Our focus includes sensor technologies that have some of their elements printed with a special interest for fully printed structures. The paper reviews the two large physical and chemical sensor families addressing different transduction principles. The paper concludes with a short notice on status and perspectives in the field with some words on the commercial maturity and trends of printed sensors on foil.

Introduction

In recent years, there has been a tremendous increase of the number of sold microsystem sensing units. Among others, Micro Electro-Mechanical Systems (MEMS)-based accelerometers, gyroscopes, magnetometers, microphones, and pressure, temperature, and humidity sensors have reached the consumer market with notably a massive penetration in mobile phones. And from the analysts and actors in the field, this is only the beginning of the emergence of a new sensor era. The establishment of the Internet of Things, the networking of smart objects, is in progress and will drive the implementation of sensors into industrial and also our daily life objects. TSensors (Trillion Sensors) movement was founded to accelerate and coordinate the development of the trillions of sensors that are expected to be needed for the development of a new generation of smart sensing systems. The latter will be deployed in several fields of applications including biomedical, energy, environment, gaming, logistics, manufacturing, personal health care, telecommunication, transport, etc.

Silicon sensors technology will be of course expanding to meet this demand but one can expect organic and printed electronics (OPE) to strongly complement silicon technology. OPE enables the development of sensors and smart sensing systems with unique characteristics such as flexibility, conformability, transparency, biocompatibility and possibility of fabrication over a large area. The materials and processes used to fabricate OPE devices are

*Corresponding author: Briand, D. (danick.briand@epfl.ch)

progressing towards greener electronics systems, being environmentally friendly, thanks to the use of additive processes and reduced infrastructure and thermal budget. Besides these unique functionalities, of high interest for better design integration into various kind of wearable and consumer products, OPE is also very attractive for the large area manufacturing at very low-cost of this extremely high number of sensing components or intelligent sensing surfaces. When dealing with such a high number of items and products, environmental impact becomes also a critical issue especially when considering their end of lifetime. OPE is here again well positioned to ensure their safe disposal, this by doing the right selection/development of materials. However, one should stay aware that in several cases, OPE sensors performances do not match those of silicon sensors, due the lack of resolution of the patterned features and to non-availability of on-chip integrated powerful Complementary Metal-Oxide-Semiconductor (CMOS) electronics for signal processing.

But there is still plenty of Research & Development (R&D) work ahead of us for the large deployment of OPE sensor technology. Indeed, fully printed sensing devices and surfaces are still at their infancy age and, in most cases, smart sensing systems are being demonstrated using a hybrid integration approach involving silicon components. Screen printing technology is already well established for the fabrication of sensors on rigid substrate such as alumina ceramic, but when looking at sensors processing on flexible polymeric and cellulosic foils, combination of lithographic and printing processes is commonly used. However, when

1369-7021/© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/4.0/). http://dx.doi.org/10.1016/ j.mattod.2015.08.001 1

RESEARCH

relevant, the tendency is being towards the implementation of an all printed approach.

In this article, we summarise the status and trends in the research and development of printed sensors on foil. Our focus includes sensor technologies that have some of their elements printed with a special interest for fully printed structures. For a detailed description concerning the working principle of the printing techniques mentioned in this paper, with particular emphasis on sensors fabrication, the Reader is invited to refer to [1]; another excellent review on printing fabrication techniques can be found in [2]. The paper reviews the two large physical and chemical sensor families addressing several transduction principles such as capacitive, resistive, amperometric, voltametric, optical, piezoresistive, piezoelectric, photoelectric, and pyroelectric. Finally we will conclude with a short notice on status and perspectives in the field with some words on the commercial maturity and trends of printed sensors on foil.

Physical printed sensors

This section reviews the current state-of-the-art of printed sensors for the detection of physical quantities, that is, temperature, position, light and mechanical stimuli (pressure and strain).

Temperature sensors

Temperature is one of the most important physical quantities, as it appears explicitly in the most basic physical laws and affects almost any typology of measurements. Sensors able to detect temperature variations and quantify them according to a previously defined scale are usually called thermometers [3].

Most of the printed thermometers presented in the scientific literature consist of metallic printed resistors and rely on the variation of resistance with temperature, defined with a known equation within the temperature range of interest [4]. Different types of materials, substrates and fabrication techniques have been utilised so far;the majority of examples reported in the literature describe resistors realised with metallic inks on flexible, plastic substrates.

Molina-Lopez et al. [5] described the fabrication of inkjet-printed silver meander-shaped resistors on flexible plastic foils (namely, polyethylene terephthalate - PET) (Fig. 1). These devices, in the considerer range of temperatures [-10 °C, +60 °C], exhibit

IflT-EPFL

FIGURE 1

Scheme of the inkjet-printed temperature sensors (Molina-Lopez et al.). The picture refers to a system composed of humidity and temperature sensors; the thermistor is the device on the top, with an inset shown on the left.

an extremely linear behaviour (coefficient of determination to a linear fit higher than 0.999) and are characterised by a Temperature Coefficient of Resistance (TCR) up to (6.52 ± 0.05) x 10-4 °C-1.

Another interesting example of printed temperature sensor was presented by Aliane et al. [6]. In this case the Authors described screen printed resistors fabricated on flexible polyethylene naphthalate (PEN) or polyethylene terephthalate (PET) foils. First of all, electrical connections were fabricated by sputter deposition of a thin layer of gold (30 nm) which was subsequently patterned by photolithography. For the fabrication of the temperature sensors, two different types of temperature-sensitive pastes were used: a Positive Temperature Coefficient (PTC) and a Negative Temperature Coefficient (NTC), the latter being based on antimony tin oxide (ATO). Printed layers had a thickness between 8 and 10 mm and were cured at 130 °C for 20 min after printing. Finally, A CYTOP-like fluoropolymer layer (~1 mm thick) was deposited on the top of the sensors as protection layer, also by screen printing. The PTC and NTC printed resistors could be used either alone or combined together within a Wheatstone bridge circuit, in order to increase the system's sensitivity. When tested alone, printed resistors exhibited a TCR of 0.05 C-1 and 0.006 C-1, respectively, for PTC and NTC pastes, within the temperature range of 20-80 °C. When used together in a Wheatstone bridge configuration, the system reached a maximum sensitivity of 0.54 V/°C at a temperature of 60 °C (input voltage Vin was kept at 48 V). Finally, the authors demonstrated the possibility of transferring their fabrication technology into a large-area fabrication process by realising 12 x 12 matrices of devices on PEN foils (38 cm x 32 cm) for thermal mapping or temperature threshold detection.

Britton et al. [7] described a procedure to fabricate nanoparti-culate silicon-based inks, starting by bulk silicon (either n-type or p-type doped) milled and reduced into thin nanoparticles powder (maximum nanoparticles diameter: 50 nm) which were subsequently suspended in ethanol; the addition of polymeric binders (such as cellulose acetate butyrate - CAB - or acrylic screen printing pastes) to the suspension was necessary to obtain the ideal viscosity for screen printing. These inks were utilised for the room temperature fabrication of thick film silicon-based resistors, screen printed on the top of low-temperature substrates such as paper;this technology was later patented [8]. These resistors are all screen printed: the silicon-based ink is printed on the top of interdigitated silver electrodes (length of electrodes: 16 mm, gap between electrodes: 0.25 mm), to obtain resistors of approximately 100 kV. These temperature sensors, tested between 20 and 60 °C, showed a negative temperature coefficient (NTC) with a beta value of 2000 ± 100 K. In 2010, PTS Sensors Ltd started the commercialisation of these sensors and very recently (March 2014) they have entered into a Purchase and Licence agreement with Thin Film Electronics ASA for the production of temperature-sensing smart labels where PTS printed resistors will be integrated within Thin Film Electronics printed radio-frequency (RF) circuits with near field communication functionality.

Infrared and light sensors

Radiation sensors are detectors able to interact with the electromagnetic radiation and produce an output signal as a response to

Materials Today • Volume 00, Number 00• September 2015

the interaction between the matter composing them and the radiation detected [9].

Radiation sensors can be classified according to the specific spectral range they are sensitive to;infrared (IR) sensors are those responding to radiation with a wavelength in the range between 700 nm and 1 mm while light sensors (sometimes simply called photodetectors) respond to electromagnetic waves whose frequency falls in the range between 380 and 700 nm.

Printing technologies have been recently used more and more often for the fabrication of photodetectors and this type of sensors, realised on large-area flexible substrates, are being currently commercialised by ISORG (France). In the majority of cases, printed photodetectors are fabricated by sandwiching the photoactive layer (usually an organic compound or blend, more seldom an inorganic layer) between two transparent electrodes. These devices can be employed for a large variety of applications; typically, but not exclusively, they are commonly used for industrial, medical, large-area sensing as well as in the field of colour sensing/recognition. A very recent example of an all-printed photodetector is provided by Aga et al. [10]. In this case, the bottom electrode is represented by a silver layer inkjet-printed on the top of a paper substrate while the active layer is an organic blend composed of solution-processable poly(3-hexylthiophene) (P3HT) and phenyl-C61-butyric acid methyl ester (PCBM), deposited by means of aerosol printing. The top electrode is made of aerosol-printed poly(3,4-ethylenedioxythiophene) doped with polystyrene sulfonate (PEDOT:PSS). Between the electrodes and the active layer, two different interlayers were also aerosol-printed in order to improve the photodetectors performances. These interlayers were PED-OT:PSS and a deoxyribonucleic acid (DNA) based biopolymer for the bottom and top electrode, respectively. The response of this photodetector has a peak in correspondence with a wavelength of 409 nm, where the responsivity reaches a value of 0.0019 A/W.

A similar device is presented by Azzellino et al. [11] but in this case the whole sensor is fabricated exclusively by inkjet-printing. The example presented by Azzellino et al. differs from other structures presented in the literature because the inkjet-printed active layer (once again, an organic blend of P3HT and PCBM) is not sandwiched between the electrodes but it is deposited on single, inkjet-printed Ag electrodes fabricated on the top of a polyethylene naphthalate (PEN) substrate. The top electrode, realised by inkjet-printing a layer of PEDOT:PSS, was laid perpendicularly to the bottom electrode thus leaving most of the active layer exposed to radiation, as shown in Fig. 2.

By carefully optimising the inkjet-printing deposition parameters, the authors were able to obtain reproducible devices with a peak External Quantum Efficiency (EQE) of 83% in correspondence with a wavelength of 525 nm.

An interesting example of ambient-light sensor was provided in 2014 by Maiellaro et al. [12]. In this paper the Authors described a sensing system exploiting an organic photodiode whose photo-current was linearly converted into an output voltage by means of a transconductance operational amplifier (in a feedback configuration). The photodiode was fabricated on a PEN flexible substrate, using a blend of P3HT and PCBM dissolved in a mixture of o-dichlorobenzene and chloroform, deposited by means of single nozzle-inkjet printing to obtain a relatively thick final active layer

(a) blend (b)

FIGURE 2

Scheme of the full inkjet-printed photodetector (a), actual picture of the device (b) (Azzellino et al.).

(~800 nm). This active layer was printed on the top of an Au electrode previously deposited on the plastic substrate and was then covered with a second metallic layer (Al, 100 nm), thermally evaporated on the organic photoactive layer. The photodiode was then encapsulated using a flexible, transparent barrier film. The operational amplifier was fabricated on a different PEN substrate, with pre-patterned gold interdigitated electrodes (IDE) (L = 20 mm and W = 2000 mm);on the top of the IDE the n-type and p-type semiconductors (Polyera ActiveInk® and TIPS-pentacene, respectively) are deposited by printing. A thin layer (750 nm) of CYTOP fluoropolymer, acting as the gate dielectric, is subsequently deposited on the organic active layer by screen printing, leaving open areas which allow access to the IDE. Finally, a 5 mm thick silver ink gate electrode is screen printed in correspondence with the underlying IDE. The two foils (the one containing the photodiode and the one containing the operational amplifier) were then connected together using a foil-to-foil assembly approach, achieved by means of an isotropic conductive glue: the cathode of the photodiode was then glued to the inverting terminal of the amplifier and the amplifier's pins were glued to a flat, flexible cable used to create the electrical connection between the sensor system and the Printed Circuit Board (PCB) used to evaluate the system's performances. This system exhibited a very good linear behaviour (R2 = 0.9946) in the illuminance range of [0;11,400] lx with a linear sensitivity of approximately 8 x 10~4 V/lx.

As for infrared sensors, Gohier et al. [13] demonstrated that inkjet-printing can be successfully used for the fabrication of IR detectors on flexible, polyimide substrates. These sensors were fabricated by inkjet-printing a water/1-propanol (4:1) dispersion of Multi-Walled Carbon Nanotubes (MWCNTs) on the top of a couple of parallel inkjet-printed silver electrodes. These resistors were then biased at a constant voltage of 0.2 V and their resistance was monitored over time while the active MWCNTs sensing area was irradiated with pulses coming from an IR source (850 nm, 0.5 mW/mm2). The sensors response could be clearly identified in terms of resistance drop when devices were exposed to IR radiation. Sensors responsivity of such inkjet-printed IR sensors was as high as 1.2 kV/W, more than four times higher than the maximum value reported in analogous devices fabricated with suspensions of MWCNTs.

Mechanical sensors

Mechanical sensors may be defined as devices able to detect and measure 'mechanical quantities' which include, but are not limited to, acceleration, pressure and strain [14].

For the sake of clarity, these three typologies of sensors will be examined separately.

RESEARCH

Accelerometers

Accelerometers are devices able to measure the acceleration experienced by a body. Nowadays, most of accelerometers consist of cantilever beams equipped with a proof mass fabricated by using the MEMS technology; these devices are typically fabricated on silicon wafers with conventional clean-room technologies. More recently, printing techniques have been employed for the fabrication of accelerometers. Hense et al. [15] reported on the fabrication of a suspended inertial mass accelerometer. A layer of PEDOT:PSS was first inkjet-printed on the top of a cyclododecane sacrificial layer, deposited onto a metallic counterelectrode. A printed capacitance-to-voltage converter was coupled with the oscillating device in order to transform the mass oscillations along the z-axis into capacitance variations which could, in turn, be converted into an output voltage signal. Improvement of the sensor's performances was obtained by glueing an additional mass of 8.77 x 10-3 g onto the printed PEDOT:PSS layer. The initial capacitance of the system was 7.5 pF, its sensitivity was 0.235 pF g-1.

Another interesting example of printed accelerometer, specifically designed for e-textile applications, is provided by Wei et al. [16]. The authors started the device fabrication by using a commercial polyester-cotton fabric as substrate, which was locally smoothened by screen printing deposition of an ultra-violet light (UV) curable planarisation layer;screen printing was also used for the deposition of the bottom electrodes, realised by using a low-temperature thermally curable silver paste (5 min at 80 °C). After deposition of the bottom electrodes, a sacrificial layer of trimethy-lolethane (TME) dissolved in cyclohexanol and propylene glycol was also screen printed; then the cantilever (made of a UV curable dielectric polymer) and the top electrode were printed. By annealing the whole structure at 160 °C, the sacrificial layer sublimated thus leaving a gap between the bottom electrodes and the cantilever. The best performances, in terms of sensitivity (0.0022 V m-1 s2), were obtained with free-standing 12 mm long and 10 mm wide cantilevers, excited at a frequency of 30 Hz (see Fig. 3a for an example of input-output curve). The same devices were also successfully tested as sensors of human movement: by wrapping the fabric containing the accelerometers around a person's forearm and connecting the sensors to a capacitance-to-voltage converter able to transmit wirelessly recorded data to a personal computer, it was possible to measure the acceleration of the person's arm, moved up and down (as shown in Fig. 3b where 'Voltage' refers to the output of the printed accelerometers and 'Acceleration' is the output of a reference accelerometer, used as comparison).

Very recently, Molina-Lopez et al. [17] reported on the fabrication of an array of printed Micro Electro-Mechanical System (MEMS) microbridges on polymeric foil. The bottom electrode of such devices consists of a printed silver layer (200 nm thin and 65 mm wide) while the top electrode/microbridge is a thicker (~2 mm) silver electrode, 80 mm wide. The gap between the two electrodes is realised by means of a photoresist sacrificial layer placed in between of them; this sacrificial layer has a thickness of approximately 5 mm, which corresponds to the height of the gap. Each functional layer is deposited by inkjet-printing technology. Though used as a humidity sensor, the system can be potentially used as accelerometer by exploiting the microbridges deflection in response to acceleration.

Pressure and strain sensors

In order to detect pressure, different transducing principles can be exploited: piezoresistive effects on strain gauges, capacitance variations of flexible structures exposed to pressure, current variations due to applied forces in Field Effect Transistors (FETs), voltage variations caused by the mechanical deformation of piezoelectric materials and so on.

The most simple one is represented perhaps by the so-called 'capacitive pressure sensors' which are typically structured in a parallel plate configuration where one of the two electrodes is fixed and the other one (where pressure is applied) is flexible. When the flexible electrode deflects under applied pressure, the gap between the two electrodes decreases, thus causing an increase of the sensor capacitance [18].

While most of the commercially available capacitive pressure sensors are fabricated using conventional CMOS technology, recently printing fabrication techniques have started being successfully used for this purpose. A first example of an all-printed capacitive pressure sensor array was presented by Narakathu et al. [19]. In this case, a flexible plastic foil (PET, 175 mm thick) was used as substrate and, on the top of it, a few silver parallel lines (0.5 cm wide and 4 cm long, with a spacing of 0.5 cm), acting as bottom electrodes, were screen printed. As a second step, the dielectric layer (polydimethylsiloxane, PDMS) was also screen printed. Gravure printing was finally chosen for the fabrication of the top electrodes, deposited on the PDMS layer and realised with the same geometry as the bottom electrodes but printed with a rotation angle of 90° in order to form the grid shown in Fig. 4a.

Each capacitor was characterised by an initial capacitance value of ~26 pF;mechanical tests performed on the sensors found out that the lowest detectable pressure had a value of 800 kPa, in correspondence of which capacitance increases of 5%. The dynamic response of the sensors was also recorded (see Fig. 4b),

Input-output curve of a screen printed accelerometer (a), comparison between a screen printed and a commercial wireless accelerometer (MicroStrain G-link) (b) (Wei et al.).

FIGURE 4

Scheme showing the full-printed capacitive array of pressure sensors (a), dynamic response of the sensors array to increasing pressure (b).

Materials Today • Volume 00, Number 00• September 2015

within a pressure range between 800 kPa and 18 MPa; capacitance increases with pressure, as expected from the thinning of the dielectric layer, with a maximum increase of 40% observed for an applied pressure of 18 MPa. The sensors also exhibited a very good reversibility, showing a return to the baseline capacitive value of 26 pF (tolerance: 0.15%) when mechanical stimuli were ceased.

Piezoelectricity is another physical phenomenon which can be exploited in order to detect mechanical stimuli. Indeed, certain materials respond to mechanical deformation by exhibiting a proportional change in their electric polarisation, thus allowing the transduction of the mechanical input into an electrical output [20]. Piezoelectric pressure sensors have been successfully fabricated for more than twenty years, the first example being the work of Morten et al. [21] describing a lead zirconate titanate (PZT) partic-ulate mixed with lead oxide binder and an organic carrier in order to obtain a screen-printable paste used to fabricate a piezoelectric diaphragm on an alumina substrate for the detection of pressure and oscillations. Much more recently, Zirkl et al. [22] demonstrated an all screen-printed array of pressure sensors fabricated on PET flexible substrates. All conductive inks used (PEDOT:PSS, carbon and silver pastes) for the fabrication of the electrodes are commercially available while the active piezoelectric layer was prepared by mixing poly(vinylidenefluoride-co-trifluoroethylene), also known as P (VDF-TrFE), pellets into a solution composed for the 50% of a low boiling point solvent (acetone) and for the remaining 50% of a high boiling point solvent (g-butyrolactone, GBL). The active layer (5 mm thick) was sandwiched between the top (fabricated with carbon paste) and bottom electrodes (fabricated with PEDOT:PSS) while connections to the read-out circuit (a capacitance-to-voltage converter) were printed with the silver paste. For the fabrication of this multi-layered structure, after each printing step a short annealing at 100 °C was needed. These printed sensors, characterised by a piezoelectric coefficient d33 between 20 and 30 pC/N, were tested in terms of response to human touch (pressure between 0 and 1.75 bar) showing an extremely linear response (R2 > 0.9945).

Other printed devices commonly employed for the detection of pressure are Organic Field Effect Transistors (OFETs). Initially used simply as addressing elements of a flexible matrix, in which the true sensors were made of pressure-sensitive rubber elements containing graphite [23], inkjet-printed OFETs were later employed as sensing elements themselves. Noguchi et al. [24] reported on the fabrication of an array of bottom-gate, bottom-contact transistors with inkjet-printed electrodes and gate dielectric, fabricated using a silver nanoparticles-based ink and a poly-imide solution, respectively. Passivation of these transistors was then achieved by the deposition of a thin layer (5 mm) of parylene-C on the top of which a suspended electrode was placed and kept constantly polarised at -100 V. The application of a force on the suspended electrode makes the electrode touch the drain and results in a measurable source-drain current increment. In another example described in the literature [25], pressure is applied directly on the channel of bottom-gate, bottom-contact transistors whose electrodes and semiconductive layer were fabricated by inkjet-printing silver-nanoparticles based ink and a solution of P3HT in toluene, respectively. The devices were then embedded between two thin layers of PDMS. These OFETs were tested by applying a

force perpendicularly to the channel in the range between 0 and 4 N, showing an almost completely reversible behaviour and a resolution of 0.05 N. The physics behind sensitivity of OFETs towards pressure is not fully understood yet, however it is speculated that several different phenomena occur when pressure is applied perpendicularly to the transistor channel (reversible deterioration of electrodes, deformation of the semiconductor crystalline structure, modification of gate dielectric thickness and structure, etc.) and contribute to the current variations shown by such devices.

Detection of mechanical stimuli can also be performed by exploiting piezoresistivity of printed materials, that is, the change in electrical resistivity produced by the application of a force which deforms the material along a certain axis [26]. It is broadly demonstrated in the scientific literature that thick film of printed ceramic/metal (cermet) composites or, more recently, polymeric materials [27] exhibit piezoresistive behaviour and can, therefore, be used in order to detect pressure and strain. One of the first examples of printed piezoresistive pressure sensors dates back to 1994 and was presented by Csaszar et al. [28]. These devices consisted of carbon-loaded polymer composite thick film resistors screen-printed on the top of a flexible, epoxy-based substrate. In the last years, several other examples of printed piezoresistors have been presented as pressure sensors. Zhang et al. [29] presented a printed piezoresistive thin structure for the detection of mechanical stimuli. In this structure, the sensing elements are represented by a matrix of circular pieces of a commercial, piezoresistive rubber composed of an insulating elastomer where a certain amount of conductive nanoparticles have been homogeneously dispersed. These sensing elements are embedded into a film of insulating silicon rubber in order to form a soft, sensing layer with a thickness of 0.9 mm. The sensing layer is then sandwiched between two electrodes; all the electrodes, as well as the conductive trace for signal transmission, are printed on the substrate layer by the screen printing technique. Variations of resistance induced by mechanical stimuli are transformed into voltage output by means of a resistance to voltage converter, digitalised and sent to a PC where they are processed with a MATLAB interface. The sensorised film was tested statically by applying forces perpendicular to its surface in the range between 0 and 9.8 N, showing a resolution of 0.15 V/N, an almost linear behaviour and hysteresis between force uploading and downloading of less than 15%. The sensors were also tested dynamically by applying force pulses on the device surfaces; this test showed a response time between 90 and 270 ms, depending on the force value. Even more recently, Thompson et al. [30] reported on the fabrication of a PEDOT:PSS based piezoresistive pressure sensor realised by means of aerosol printing technique. The sensors presented in this paper were meander-shaped resistors, fabricated on a polyimide (PI) Kapton flexible substrate. Resistance variations due to strain were recorded by inserting the piezoresistors inside a Wheatstone bridge circuit and measuring the corresponding voltage difference between two midpoints of the circuit. The sensing protocol consisted of applying a series of 3000 sinusoidal strain pulses (peak amplitude of 2000 me and frequency of 0.5, 1 and 2 Hz) to the sensors and recording the corresponding resistance relative variation (AR/R0). Sensors were able to rapidly follow the applied strain (response time less than 1 ms, regardless of the frequency) and showed a

RESEARCH

peak response in the range 0.1-0.15%;the maximum gauge factor (0.53) was obtained in correspondence with an excitation frequency of 2 Hz. A very recent example of bio-inspired printed strain sensor was published by Harada et al. [31]. In this paper the Authors present an array of printed sensors shaped as mammalian whiskers (e-whiskers) for the detection of strain. These sensors were fabricated on a flexible silicone rubber substrate, using a mixed paste composed of a Single-Walled Carbon Nanotubes (SWCNTs) dispersion and a silver nanoparticles-based conductive ink; the paste was deposited on the substrate by means of screen printing, obtaining a dried final layer with thickness of ~1.5 mm. Devices were fabricated as single-meander resistors, with fixed electrodes length (Lsensor). Finally, the silicone substrate was patterned by laser cutting, to isolate each resistor from the surrounding devices and to form the e-whiskers array. Devices were tested by measuring the normalised resistance values R/R0 as a function of the whiskers' displacement caused by the application of a force perpendicular to the whiskers' tip (displacement range of [-3;+3] mm, 0 mm being assumed as the sensors' position at rest, in correspondence of which the measured resistance was R0). The different fabrication parameters (silicone substrate thickness, SWCNTs and Ag nanoparticles weight ratio and Lsensor) were varied, in order to maximise the sensors' sensitivity; the best performances were found for a silicone substrate 1.5 mm thick, a SWCNTs:Ag nanoparticles weight ratio of 10:8 and Lsensor measuring 3 mm. Under these conditions, devices showed a linear response within a pressure range of ~±450 Pa, in correspondence of which the induced strain caused a normalised resistance variation of 4500%, thus leading to a maximum linear sensitivity of 59%/Pa, more than seven times higher than values previously reported in the scientific literature. The response time was also evaluated, by applying forces able to cause whiskers' deflection of -0.5 mm (tensile stress) or +0.5 mm (compres-sive stress). In both cases, response time was estimated to be approximately 90 ms.

Chemical printed sensors

This section reviews the current state-of-the-art of printed sensors for the detection of chemical analytes, that is, biomolecules, ions and gaseous species.

Biosensors

Biosensors are devices able to monitor biomolecular interactions in real time. In biosensors, one component (usually called 'ligand' or 'receptor', which can be either a biological or a non-biological molecule) is immobilised on a solid surface while the component to be detected (the 'analyte', by definition a biological molecule) is dispersed into a solution put in touch with the surface. The interaction between the ligand and the analyte results in a change of the electrical properties of the solid surface which can be measured in order to determine the amount and typology of analyte that has reacted with the ligand [32].

Typically, but not exclusively, biosensors are electrochemical systems where the ligand is immobilised on the working electrode; reactions between the ligand and the analyte cause variations of the current measured between the counter and reference electrodes and such variations can be used to determine if and how much analyte is present into a certain test solution.

Printing techniques have been successfully employed for the fabrication of biosensors, in particular if the biosensor is represented by an electrochemical transducer where the electrodes can be easily fabricated by means of printing methods. Among the different printing techniques, screen printing is by far the best established technique to fabricate electrochemical biosensors, because of its low cost and the possibility to reach mass production [33]. So far, screen printing has been used for the fabrication of a large variety of electrodes (gold, platinum, silver, palladium, graphite) fabricated over several different substrates (alumina, ceramics, polyvinyl chloride - PVC, fibre glass) [34].

While screen printing remains the most important technique for the fabrication of biosensors electrodes, more recently inkjet-printing has also been used for the fabrication of biosensors. One of the first examples of inkjet-printed biosensors was provided by Jensen et al. [35]. The authors reported on the fabrication of an amperometric sensor for the detection of a cancer biomarker (namely, interleukin-6 (IL-6)) in serum. Devices were fabricated on a flexible polyimide substrate; they consisted of an array of 8 gold working electrodes (465 mm x 465 mm) inkjet-printed directly on the substrate using a homemade Au nanoparticles-based ink. After fabrication of the electrodes, a passivating layer consisting of a homemade poly(amic acid) ink was also inkjet-printed on the top of the devices, leaving just small openings in correspondence with the working electrodes' active areas and electrical contacts. A single array fabricated as described by the Authors has very limited production costs (less than 0.2 €). After fabrication, electrodes were functionalised using a 4 mM 3-mercaptopropionic acid (MPA) solution in ethanol, to form a self-assembled monolayer (SAM) on the top of the working electrodes; in a second phase, biotinylated secondary antibodies with 16-18 horseradish peroxi-dase labels were attached to the SAM carboxylic groups. Finally, the calf serum containing the markers to be detected was deposited on the electrodes and incubated at room temperature for 80 min. Sensors were characterised by a linear response (R2 = 0.99837) within the range 20-400 pg mL-1 (the range of interest for the diagnosis of many types of cancer) with a detection limit of 20 pg mL-1; within this range, sensitivity was 11.4 nA pg-1 cm-2. More recently, Lesch et al. [36] presented an example of all inkjet-printed biosensor fabricated on Kapton flexible sheets for the detection of antioxidants (such as ascorbic acid, AA) in biological samples. In this case, an electrode is fabricated by inkjet-printing a silver nanoparticles-based ink then the sensor's active area (500 mm x 500 mm) is realised by inkjet-printing on the top of the silver electrode a Double-Walled Carbon Nanotubes (DWCNTs) dispersion. In order to avoid direct contact between the aqueous solutions and the silver electrode, a final UV-curable insulating layer is inkjet-printed all around the active area and on the top of silver electrode (see Fig. 5 for an optical image of the final device).

The system was successfully tested in order to detect the anti-oxidant power of two physiological solutions, 0.5 mM and 1 mM AA, obtaining 128 nW and 253 nW, respectively (ratio: 1.97), in good accordance with the concentration ratio of the two solutions (2). The same sensors were also used in order to measure the concentrations of antioxidant species (uric acid, glutathione and so on) into blood samples taken from volunteers.

Another recent example is the one provided by Ihalainen et al. [37]. In this paper, the Authors presented an impedimetric system

Materials Today • Volume 00, Number 00• September 2015

FIGURE 5

Inkjet-printed ascorbic acid sensor developed by Lesch et al.

for the detection of DNA hybridisation, fabricated on a disposable paper substrate. The system consisted of gold electrodes fabricated by inkjet-printing (thickness 550 ± 100 nm, active area either 0.28 or 0.08 cm2);these electrodes were then used in order to fabricate two different types of recognition architectures. The first type consists of alternate layers of a binary mixture of biotinylated hexa(ethyle-neglycol) undecanethiol (Biotin-PEG-thiol) and 11-mercapto-1-undecanol (MUOH) (85:15 mol%), self-assembled monolayer (SAM), streptavidin (SA) and biotinylated DNA probe (biotin-DNA probe). The other recognition architecture is based on a mixed SAM consisting of a thiol-functionalised DNA probe (HS-DNA probe) and 11-mercapto-1-undecanol (MUOH). Surface Plasmon Resonance (SPR) was used to examine the binding capacity and selectivity in both systems and it was shown that the HS-DNA/MUOH system exhibited a higher binding capacity for the complementary DNA target. After exposure to complementary and non-complementary DNA target analytes, the sensors' response was characterised by means of Electrical Impedance Spectroscopy (EIS), varying the input signal frequency between 1 Hz and 1 MHz. While exposure to non-complementary DNA target causes almost no effects on the impedance response of the sensors, exposure to complementary DNA sequences is responsible for a marked increase of the impedance capacitive component in both recognition architectures, with a relative capacitance increase peak of ~20% (at 3.5 kHz) for the MUOH:Biotin-PEG-thiol SAM-SA-biotin-DNA probe system.

Alongside screen and inkjet-printing, another technique which has received much attention is gravure printing, especially in those applications where a higher printing resolution (for instance for the fabrication of interdigitated electrodes) is needed [38]. For instance, Narakathu et al. recently reported on the successful employment of gravure printing for the fabrication of a silver electrochemical biosensor on PET substrate for the detection of toxic biomolecules (such as D-proline and sarcosine, involved in the degradation of collagenous and muscular tissues, respectively) in biological fluids [39]. This biosensor, occupying an area of 1 cm x 2 cm, is characterised by a circular inner working electrode with a diameter of 1.7 mm and was successfully used for the detection of analytes concentration down to 1 pM.

biology [41], in medicine and health care [42], in food [43] and chemical/materials science industry [44].

Printing techniques have proved to be ideal candidates for the fabrication of pH sensors which are portable, robust, sensitive and cheap. Most of the sensors or sensing systems described in the scientific literature are realised by employing screen printing as a fabrication technique for the electrodes of the pH sensor. Kam-pouris et al. [45] reported on the realisation of an all-screen-printed pH sensing platform where electrodes were fabricated by sequentially depositing carbongraphite, silver-silver chloride and dielectric inks onto a polyester substrate material. The reference electrode was fabricated by mixing the carbongraphite ink with 1,1-dimethylferrocene while the working electrode was realised by mixing the carbongraphite ink with 9,10-phenanthraquinone. With such system, the Authors were able to successfully measure the pH of aqueous solutions in a large range (1-13), with a sensitivity of 57 mV per pH unit and excellent reproducibility (95% confidence).

Even if screen printing still remains the most frequently utilised printing technique for the fabrication of pH sensors on flexible substrates, some reports on inkjet-printed sensors have recently started to appear in the scientific literature. Maattanen et al. [46] were very recently able to fabricate an inkjet-printed three-electrode system on a coated paper substrate (see Fig. 6a for a device scheme). The working electrode (WE) and counter electrode (CE) were realised by using gold nanoparticles-based inks while the quasi-reference electrode (QRE) was fabricated by inkjet-printing a silver nanoparticles-based ink and was, in a second step, electro-chemically modified and transformed into an Ag/AgCl electrode through a chlorification process. On the WE surface, a layer of polyaniline (PANI) was electropolymerised and used as a potenti-ometric pH sensing layer. The sensor active area was then defined by means of a printed hydrophobic PDMS layer, leaving a part of the electrodes exposed to the aqueous test solutions. This printed sensing system was tested in a pH range between 2 and 10 and showed sensitivity close to -59 mV per pH unit (see Fig. 6b for the calibration curve). Even more recently, Song et al. [47] described another inkjet-printed system for the detection of pH fabricated on common, flexible transparency films (145 mm thick). These sensors are chemiresistors whose electrodes are realised by inkjet-printing several layer of a conductive Multi-Walled Carbon Nano-tubes (MWCNTs)-based ink while the active layer was obtained by filling the gap between the two electrodes with an inkjet-printed polyaniline (PANI) layer, obtained starting from a homemade water-based ink containing PANI nanowires (few mm in length,

pH sensors

pH sensors are devices able to detect the activity of hydrogen ions into an aqueous solution. The most common typology of pH sensor is made up of a couple of electrodes: the first electrode, called reference electrode, is characterised by a very stable half-cell potential while the potential of the second electrode, called pH electrode, depends on the concentration of hydrogen ions into the solution. The two electrodes are connected to a voltmeter able to measure the potential difference between them which, in turn, is used to derive the pH value [40].

pH sensors are characterised by an incredibly wide spectrum of applications; for instance, they are commonly employed in

FIGURE 6

(a) Scheme of the inkjet-printed pH sensor. (b) Sensor response to pH variations (Maattanen et al.).

RESEARCH

diameter between 100 and 150 nm). The sensors were initialised by exposing them to a strong acid, in order to maximise the output current (I0), then they were immersed into solutions at different pH and current variations were measured keeping the voltage constant. These devices were successfully tested in the pH range [1-6] by measuring the normalised current response I/I0 as a function of pH. As expected, when pH increases the conductivity of the PANI layer decreases because of deprotonation;this is clearly reflected by a drop in the normalised current response from pH 1 (I/I0 = 1) to pH 6 (I/I0 - 0.4).

Ion sensors

Ionic sensors maybe defined as devices able to convert the activity of a specific ion dissolved into a certain solution into a measurable electric potential [48]. They are mostly used to detect and quantify the concentration of a certain ionic species into a given solution, with a very wide spectrum of potential applications, varying from the detection of certain toxic metallic cations in biological fluids to water quality control.

The majority of printed ionic sensors presented in the scientific literature are based on the Ion Selective Electrode (ISE) working principle, even though more recently also Organic Thin Film Transistors (OTFTs) have been proposed for the detection of ionic species [49].

Historically, screen printing was the first printing technique to be explored for the fabrication of silver electrodes [50,51] or even ion-selective membranes [52] used in ISE-based systems. Today, screen printing still remains the most important fabrication technique for the production of silver [50], gold [53], carbon [54] and copper sulphides [55] ISE, just to mention a few examples. A recent example of an all-screen printed ion sensor for the detection of trace lead is provided by Laschi et al. [56]. In this paper, the authors describe a screen printed electrochemical system fabricated on flexible, polyester substrates; commercial silver-based paste (Elec-trodag PF-410) was used for the fabrication of the conductive tracks as well as for the pseudo-reference electrode while the auxiliary electrode was fabricated with a commercial graphite-based paste (Electrodag 423 SS). The working electrode was screen printed using a gold-based paste (R-464 from Ercon Inc.) and finally, an insulating layer (Vinylfast 36-10) was screen printed on the top of the interconnections in order to avoid short circuits between the electrodes and the test solution. The system was successfully used in order to detect Pb(II) ions into river water samples (diluted 1:10 in deionised water and acidified with a 0.1 m HCl solution), exhibiting a sensitivity of 0.08 mA/mgl-1, a detection limit of 0.5 mgl-1 and a linear behaviour in the range 050 mg l-1. More recently, inkjet-printing has started to be employed in the field of ion sensing, for the deposition of the ion-sensitive layer. Crowley et al. [57] reported on the fabrication of a sensor based on an inkjet-printed polyanaline nanoparticle-modified electrode for the detection of ammonia in aqueous solutions (more specifically, for the detection of the ammonium ion NH4+). This system was composed of a silver wire (0.1 mm of diameter), used as reference electrode, a platinum wire (0.5 mm of diameter) used as auxiliary electrode and an inkjet-printed working electrode fabricated with an aqueous solution of polyaniline (PANI) nanoparticles doped with dodecylbenzene sulphonic acid (DBSA). All the necessary electrical connections were realised by

FIGURE 7

Current vs concentration response of a sensor described by Crowley et al. Calibration curve (left) and real time response (right).

screen printing a carbon conductive paste. These sensors proved to be able to detect ammonium ions in water solutions and were characterised by an extremely linear response within the range 080 mM (which corresponds to a concentration range of 01.44 ppm). Linear sensitivity was of 2 x 10-8A mM-1 while the detection limit was 2.58 mM (0.44 ppm) (see Fig. 7 for a typical sensor response).

Gas and vapour sensors

Since Mid-2000s, there have been several works performed on the fabrication of chemical vapour/gas sensors on polymeric and cellulosic substrates. First developments were mainly focused on humidity sensors and organic field-effect transistors for the detection of vapour and gases using standard thin film processing techniques such as spin coating and photolithography. A review on these devices has been published by Briand et al. in 2011 [58]. Since then several works have been addressing the printing of the gas sensitive films, the transducers and even of the whole sensors structure. Screen printing technique is an established process to print electrochemical gas sensors on ceramic substrate [59]. A well-known device is the potentiometric Yttria Stabilised Zirconia (YSZ) oxygen sensor, commercially available under the name of lambda sensor, which operates at high temperature. These sensors measure the electrical potential of an electrode when no current is flowing. Its two electrodes provide an output voltage corresponding to the quantity of oxygen in the exhaust relative to that in the atmosphere. On the printing of electrochemical and chemoresistive gas sensing devices on polymeric foil, the conducting polymer poly-aniline (PANI) has been significantly studied [60,61]. In [62] is provided a review of work performed on printing conducting polymers mainly for chemical sensors development. Crowley et al. reported on a fully printed polyaniline-based gas sensor for the detection of hydrogen sulphide with a detection limit of 2.5 ppmv with a linear relationship between measured current and concentration over the 10-100 ppmv region [63]. Polyaniline and copper (II) chloride layers were inkjet-printed on screen printed silver intergitated electrodes (IDE) on a flexible PET substrate. Using the same types of materials but different printing methods, hydrogen sulphide gas sensor was demonstrated on paper substrate. The sensing film was produced by screen printing and spray coating the PANI-CuCl2 on IDE inkjet-printed on the paper substrate. It was claimed that the large roughness and porosity of the paper substrate offers and increased surface sensing area. Also on paper, H2S sensors made of inkjet-printed copper acetate (CuAc) with silver interdigitated electrodes have been produced and extensively characterised [64]. The same group has also reported

Materials Today • Volume 00, Number 00• September 2015

on the use of copper acetate (CuAc) as simple optical indicator. Integration of the sensor in a printed circuit consisting of an electro-chromic pixel connected in series with a 1.5 V printed battery was also demonstrated. In [65], an improved CuAc-based chemiresistor-type sensor capable of a fast response time and sub-ppm sensitivity towards H2S gas at room temperature is presented. These improvements in sensing properties were achieved by the oxidative conversion of solid-state CuAc nano-particle to Cu2O (and its cross defect structure Cu3O2) nanocrystals using plasma treatment and by incorporation of gold nanoparticles (Au NP) in the printed film. In 2013, breath ammonia sensors based on printed polyaniline nanoparticle sensors operating in the sub-ppm concentration range have been reported [66]. A simple fabrication of paper-based flexible ammonia gas (NH3) sensor with silver and poly(m-aminobenzene sulphonic acid) functionalised Single-Walled Carbon Nanotubes (SWNT-PABS) via inkjet-printing was reported in [67]. Silver dispersion was first inkjet-printed onto the photopaper to prepare the electrodes. SWNT-PABS dispersion was then printed to form the gas sensitive layer. The paper-based sensor showed excellent sensor response, short response and recovery time to different concentrations of NH3 at ppm level, and could be stable for several months [67]. A flexible and lightweight chemiresistor for NO2 detection was printed on polyamide using the gravure technique for both the silver electrodes and the WO3-PEDOT:PSS [68]. Detection of 50 ppb of NO2 was demonstrated at room temperature with a response time below the minute and a recovery time slightly below 90 s. Gravure printing was applied in that work, but the resolution of the structures could be achieved with any other printing techniques, such as inkjet printing. Gravure might be of interest only if high volume and production throughput are required. Stability of such an architecture based on silver electrodes as transducer needs also to be confirmed. Printed electrochemical gas sensors based on ionic liquid electrolytes (ILE) have also been investigated with interesting results. Amperometric gas sensors on plastic substrate based on screen printing and lamination strategies to form a small, low power (mW) sensor package is presented in [69]. The sensors include three electrode cells comprising a printed working electrode for gas detection and a Pt auxiliary and quasi-reference electrode. CO sensors were prepared using acidic electrolytes such as 4 M H2SO4 while NH3 sensors were constructed investigating different room temperature ILEs. Another example is a high-sensitive and flexible paper-like electrochemical oxygen sensors based on ILEs (BMIMPF6) made by printing nanoporous gold electrode arrays on cellulose membrane [70]. The sensor looked like a piece of paper but possessed high sensitivity for O2 in a linear range from 0.054 to 0.177 v/v%, along with a low detection limit of 0.0075% and a short response time of less than 10 s. Printing was also applied to the fabrication of humidity sensors using a resistive or capacitive transduction principle. Yun et al. [71] reported on transducers made of screen printed silver paste on polyimide with electroless plating of Cu/ Ni/Au with a humidity sensing polyelectrolyte ink crosslinked and anchored to the gold electrode. Humidity-sensing membranes were composed of copolymers of [2-(methacryloyloxy)ethyl] dimethyl propyl ammonium bromide (MEPAB), PVPEM and methyl-methacrylate, followed by thiolation. The humidity sensor exhibited a good sensitivity changing the resistance approximately four orders of magnitude for relative humidity range varying

from 20 to 95% R.H., a fast response and recovery time, good linearity, and small hysteresis, with a precision of ±0.9% R.H. Targeting low power consumption, the development of capacitive humidity and vapour sensors were also studied. These sensors utilise the ability of certain polymers to absorb humidity and other vapours and gases, modifying their dielectric properties and/or swelling, inducing in that way changes in the capacitive structure in which they are integrated. A gravure-printed silver electrodes on PET foil covered with the polymer poly (2-hydro-xyethyl methacrylate) (pHEMA) for capacitive humidity sensing was presented in [72]. The sensor response was measured in the range of 30% R.H. to 80% R.H.;the maximum percentage change in capacitance was 172% at 80% R.H. when compared to base capacitance at 30% R.H. All the other examples to be described here were made using inkjet printing of silver planar IDE. One group used Nafion® as humidity sensing layer on polyimide substrate [73]. The sensors were tested in a broad humidity range of 5-95% R.H. at frequency varying from 20 Hz to 1 MHz exhibiting a non-linear response. Another work by Rivadeneyra et al. [74], using directly the polyimide substrate as sensing layer, achieved sensitivity of 4.5 fF/% R.H. A thermal coefficient of -0.4 fF/C was measured at 100 kHz, whereas the sensitivity and the thermal coefficient were 4.2 fF/% R.H. and -0.21 fF/C, respectively, at 1 MHz. This latter result implies that thermal compensation could be avoided depending on the required accuracy and the chosen frequency. The performance of the previous printed capacitive humidity sensors was limited and no special care was taken to study and remove the parasitic effect of the substrate in the majority of the cases [75]. This usually resulted in long or no saturation of the sensor signal resulting in slow response time. Transient behaviour of the printed humidity sensors was studied by Molina Lopez et al. [75]. Molina-Lopez developed different kinds of printed capacitive vapour sensors, namely interdigitated electrode (IDE) with a sensing layer on top [76], and parallel-plate (PP) sensors where the sensing layer is placed between plates [77]. Sensors were fabricated by inkjet-printing of conductive silver electrodes (optionally passivated with Ni) on PET substrate with a cellulose acetate butyrate (CAB) humidity sensing layer. The standard PP structure outperformed the others investigated in terms of capacitance per surface area (3.8pF/mm2), sensitivity towards R.H. (0.33%/% R.H.) and response time (50 s). The high porosity of the printed top electrode facilitated the humidity diffusion through the sensor, permitting the realisation of devices as small as 1 mm2. The area and the response of the sensor could be even decreased further using a micro-contact printing process to pattern the top electrode [78]. Molina Lopez et al. proposed as well the simultaneous printing of a temperature sensor (RTD with TCR of 1.82 x 10-3 °C-1) made of nickel electroplated inkjet-printed silver for thermal compensation of the humidity sensor signal and their subsequent encapsulation at foil level [5] (Fig. 8).

Printed heating structures have also been reported to heat up for improved gas sensing performances in the case of some gas sensing layers. Claramunt and colleagues have reported an ammonia sensor consisting of a polyimide flexible substrate, onto which an array of silver interdigitated electrodes was inkjet-printed on one side, and a common silver heater on the backside [79] (Fig. 9). The interdigitated electrodes were coated with carbon nanofibers decorated with metal nanoparticles. The heater was used to

RESEARCH

FIGURE 8

Capacitive humidity and resistive temperature sensors made using inkjet printing and encapsulated at the foil level (from [5]).

FIGURE 9

Optical picture of the sensor developed in [79]. A silver heater was printed on the bottom part of a polyimide foil and four comb electrodes on the other side of the foil to be functionalised with gas sensing layers.

chemically reset the nanofibers and make the sensor operation reversible. A relatively high power consumption of around 150 mW was needed to achieve the optimal sensing temperature of 110-120 0C.

In Danesh et al. [80], another sensor configuration was utilised to decrease the power consumption of the device. The 1 mm2 micro-hotplate was made of inkjet-printed silver on polyethylene naphthalate (PEN) foil for operation up to 100 0C (Fig. 10a). The

FIGURE 10

(a) Schematic view of the printed micro-hotplate type gas sensor. (b) Response to ammonia of the sensor coated with a PANI layer at room temperature and when heated up at 80 °C for better signal recovery (from [80]).

(a) Schematic of a colorimetric gas sensor using the plastic foil as waveguide. (b) Optical picture of the polymeric optical waveguide inkjet-printed ammonia sensitive colorimetric film (from [85]).

optical with

microheater was separated from the sensing IDE by a laminated thin insulating organic layer. Improved long-term stability of the sensors could be enabled by electrodeposition of nickel on the heater or gold on the silver IDE. The IDE was used then as a chemiresistor by placing on top a layer of polyaniline (PANI) doped with poly(4-stryenesulphonic acid) that was synthesised and deposited on the IDE by vapour-phase deposition The proper operation of the PANI layer to detect ammonia in air required high temperature of 80 to 100 °C for a reduced recovery time (Fig. 10b). Printed gold hotplates on polymide foil have also been reported, enabling a higher temperature of operation compatible with metal-oxide sensing films [81]. A fully printed, meaning sensing layer and transducer, SnO2 gas sensor was demonstrated by Rieu et al. [81].

Colorimetric sensors offer some advantages compared to more conventional principles (resistive, capacitive) such as it can be very selective towards a specific targeted gas. The transduction principle of a colorimetric sensor is based on the change in light absorption of a chemochromic material. A shift of its maximum absorption wavelength (i.e. a change in colour) occurs when the film is exposed to the analyte to be detected. A large variety of gases may be detected by this technique. Colorimetric films for the detection of different types of gases such NH3, CO and NOx, have been reported [82-84] with limit of detections in the ppm and the sub-ppm range. Specific colour dyes are included in a polymeric film matrix. The successful local printing of these films by inkjet on plastic substrate, used as optical waveguide [85], is of interest for the formation of sensors arrays (Fig. 11).

Finally, there are several works focusing on the printing of gas sensitive layers themselves. Several polymeric materials can be deposited by inkjet printing. Inkjet-printing of CNTs was applied for the detection of NO2 and H2S using a printed RF antenna [86], and resistive and FET transducers [87], respectively. We will cite here a few examples of metal-oxide gas sensitive layers that were successfully patterned using inkjet printing. One example is surface-modified WO3 nanoparticles printed on silicon substrate and post-cured with heating cycles up to 500 K. Nanoparticles characteristics and gas sensing behaviour were evaluated for the efficient discrimination of various gaseous analytes (NO, CO, H2, H2S), from sub-ppm up to nearly 0.1% concentration levels in air [88]. Shen [89] inkjet-printed pure and doped SnO2 films on silicon and alumina substrates. The use of inkjet-printing allowed an easily doping by consecutively printing SnO2 and a dopant. His printed layers however required an annealing at 550 °C, making them incompatible with plastic substrates. This drawback was counteracted by Peter et al. [90], who developed a titanium-doped chromium oxide (CTO) ink that did not require any firing.

Materials Today • Volume 00, Number 00• September 2015

However, the adhesion to silicon substrate and the film stability were improved by sintering the printed layer at 400 °C. This temperature is however compatible with high performance polymer such as some polyimides. Finally, being an additive technique, inkjet-printing is of high interest to locally pattern different sensing films on one foil for array formation.

Conclusions and future perspectives

We have seen through this paper that there is a large set of printed sensor technologies on polymeric foil that is being researched and developed. However, their level of maturity (TRL) strongly differs, with few of them being already commercial products, with some having their production established, and many others being more less advanced lab prototypes, at the early stage of their development, or at the proof of concept level.

The most advanced printed sensor products include glucose test strips, electrocardiography (ECG) electrodes, pressure, strain and touch sensors, and photodetectors. Physical mechanical silicon sensors, for example, accelemeters, gyroscope and pressure sensors, relying on movable structures, harder to precisely micro-structured on foil, have become highly performing with the new product generations driven more by sensors fusion, compactness and cost reduction. However, the sector of chemical and biological sensors is far less mature and printed and foil technologies are surely complementary but also competitive to silicon technology. OPE sensor technology is of high interest for inclusion in cost-effective diagnostic and environmental monitoring systems, such as in wireless smart labels, micro-fluidic and lab-on-chip systems. The additive and local character of their processing is beneficial for the multilayer functionalisation of chemical and biological sensor systems.

We can notice that there are increasing efforts to develop fully printed sensor solutions but its necessity depends on the targeted applications and markets. With the large volumes expected in the near future, large area manufacturing at low-cost based on additive and printing processes will receive more and more attention. Nevertheless, several issues related to very high volume production still require to be addressed, such as sensors testing and calibration, and in some cases, their specific encapsulation, which is strongly product and application dependant. These steps can represent a large part of the sensor production cost and should not be underestimated. Besides foil based, stretchable sensor technologies for compliant and human-centred design applications will be growing in the coming years and their processing using additive and large area compatible techniques, such as printing, proposes interesting challenges to overcome. Other topics which were not addressed at all in this article and deserves consideration is the application of 3D-printing to sensors making and the transfer/ integration of nanostructures and two-dimensional materials, such as grapheme, on foil substrates.

Finally, the sensor component alone is not a solution and their deployment will depend on how successful their integration into systems will be. For more computing demanding applications, the first systems will be mainly based on a co-integration approach with silicon components. But if successful, all organic and printed electronics systems including electronics, memory, power supply, and communication will enable a new generation of smart objects and large area intelligent surfaces for applications we cannot even think about at the moment.

References

[1] S. Khan, L. Lorenzelli, R.S. Dahiya, IEEE Sens. J. 15 (2015) 3164-3185.

[2] C. Ru, et al. J. Micromech. Microeng. 24 (2014) 053001.

[3] J. Scholz, T. Ricolfi, Sensors, Thermal Sensors, John Wiley & Sons, USA, 2008. p. 13.

[4] S. Ghosh, Fundamental of Electrical and Electronics Engineering, PHI Learning Pvt. Ltd, India, 2007. p. 9.

[5] F. Molina-Lopez, et al. J. Micromech. Microeng. 23 (2013) 025012.

[6] A. Aliane, et al. Microelectr. J. 45 (2014) 1621-1626.

[7] D.T. Britton, M. Harting, Pure Appl. Chem. 78 (2006) 1723-1739.

[8] D.T. Britton, M. Harting, Printed Temperature Sensors, Patent Application Publication, US 2013/0203201, 2013.

[9] N. Tsoulfanidis, S. Landsberger, Measurements and Detection of Radiation, 3rd ed., CRC Press, USA, 2010.

[10] R.S. Aga, et al. IEEE Photonic Technol. L 26 (2014) 305-308.

[11] G. Azzellino, et al. Adv. Mater. 25 (2013) 6829-6833.

[12] G. Maiellaro, et al. IEEE Trans. Circuits Syst. - I 61 (2014) 1036-1043.

[13] A. Gohier, et al. Appl. Phys. Lett. 98 (2011) 063103.

[14] H. Bau, Sensors Mechanical Sensors, John Wiley & Sons, USA, 2008.

[15] A. Hense, et al. Proc. Eng. 5 (2010) 713-716.

[16] Y. Wei, et al. Meas. Sci. Technol. 24 (2013) 075104.

[17] F. Molina-Lopez, D. Briand, N.F. de Rooij, MEMS 2014, 2014, 506-509.

[18] S. Beeby, MEMS Mechanical Sensors, Artech House, USA, 2004.

[19] B.B. Narakathu, et al. IEEE Sensors 2012, 2012, 1-4.

[20] G. Gautschi, Piezoelectric Sensors: Force Strain Pressure Acceleration and Acoustic Emission Sensors Materials and Amplifiers, Springer, Berlin Heidelberg/Germany, 2011.

[21] B. Morten, G. De Cicco, M. Prudenziati, Sens. Actuators A: Phys. 31 (1992) 153-158.

[22] M. Zirkl, et al. Proc. SPIE 8831 (2014) 883124.

[23] T. Someya, et al. PNAS 102 (2005) 12321-12325.

[24] Y. Noguchi, T. Sekitani, T. Someya, Appl. Phys. Lett. 89 (2006) 253507.

[25] P. Cosseddu, et al. IEEE BioRob, 2012, 1907-1912.

[26] D.D.L. Chung, Composite Materials: Science and Applications, Springer Science & Business Media, Germany, 2010.

[27] K.I. Arshak, et al. SPIE 2780 (1996) 2-8.

[28] C. Csaszar, G. Harsanyi, Sens. Actuators A: Phys. 41-42 (1994) 417-420.

[29] X. Zhang, Y. Zhao, X. Zhang, Sens. Rev. 32 (2012) 273-279.

[30] B. Thompson, H. Yoon, IEEE Sens. J. 13 (2013) 4256-4263.

[31] S. Harada, et al. ACS Nano 8 (2014) 3291-3927.

[32] A. Sadana, Biosensors: Kinetics of Binding and Dissociations Using Fractals, Elsevier, The Netherlands, 2003.

[33] J. Gonzalo-Ruiz, et al. Biosens. Bioelectron. 22 (2007) 1517-1521.

[34] M. Tudorache, C. Bala, Anal. Bioanal. Chem. 388 (2007) 567-578.

[35] G.C. Jensen, et al. Phys. Chem. Chem. Phys. 13 (2011) 4888-4894.

[36] A. Lesch, et al. J. Electroanal. Chem. 717-718 (2014) 61-68.

[37] P. Ihalainen, et al. Nanotechnology 25 (2014) 094009.

[38] A.S.G. Reddy, et al. Proc. Eng. 25 (2011) 956-959.

[39] B.B. Narakathu, et al. in: Proceedings of IEEE Sensors Conference 2011, 2011, pp. 577-580.

[40] X. Zhang, et al., Electrochemical Sensors, Biosensors and their Biomedical Applications, Academic Press, USA, 2011.

[41] W. Khramtsov, et al. Cell. Mol. Biol. Lett. 46 (2000) 1361-1374.

[42] P.A. Oberg, et al., Sensors Applications, Sensors in Medicine and Health Care, John Wiley & Sons, USA, 2006.

[43] C. Bohnke, H. Duroy, J.L. Fourquet, Sens. Actuators B: Chem. 89 (2003) 240-247.

[44] K. Lee, S.A. Asher, J. Am. Chem. Soc. 122 (2000) 9534-9537.

[45] D.K. Kampouris, et al. Anal. Methods 1 (2009) 25-28.

[46] A. Määttänen, et al. Sens. Actuators B: Chem. 177 (2013) 153-162.

[47] E. Song, et al. Microelectron. Eng. 145 (2015) 143-148.

[48] F. Banica, Chemical Sensors and Biosensors: Fundamentals and Applications, John Wiley & Sons, USA, 2012.

[49] J.T. Mabeck, G.G. Malliaras, Anal. Bioanal. Chem. 384 (2006) 343-353.

[50] G.S. Cha, et al. Anal. Chem. 63 (1991) 1666-1672.

[51] H.D. Goldberg, et al. Sens. Actuators B: Chem. 21 (1994) 171-183.

[52] D.V. Chernyshov, et al. Mendeleev Commun. 18 (2008) 88-89.

[53] M. Vazquez, et al. Sens. Actuators B: Chem. 97 (2004) 182-189.

[54] G.G. Mohamed, et al. Anal. Chim. Acta 673 (2010) 79-87.

[55] R. Koncki, et al. Fresenius J. Anal. Chem. 367 (2000) 393-395.

[56] S. Laschi, I. Palchetti, M. Mascini, Sens. Actuators B: Chem. 114 (2006) 460-465.

[57] K. Crowley, et al. Analyst 133 (2008) 391-399.

[58] D. Briand, et al. Mater. Today 14 (2011) 416-423.

[59] C. Lucat, F. Menil, H. Debeda, Printed Gas Sensors Based on Electrolytes, Wood-head Publishing Limited, India, 2012.

RESEARCH

[60] L. Crowley, et al. Chem. Pap. 67 (2013) 771-780.

[61] J. Sarfraz, et al. Thin Solid Films 534 (2013) 621-628.

[62] B. Weng, et al. Analyst 135 (2010) 2779-2789.

[63] K. Crowley, et al. IEEE Sens. J. 10 (2010) 1419-1426.

[64] J. Sarfraz, et al. Sens. Actuators B: Chem. 191 (2014) 821-827.

[65] J. Sarfraz, et al. RSC Adv. 5 (2015) 13525-13529.

[66] T. Hibbard, et al. Anal. Chem. 85 (2013) 12158-12165.

[67] L. Huang, et al. Sens. Actuators B: Chem. 197 (2014) 308-313.

[68] Y. Lin, et al. Sens. Actuators B: Chem. 216 (2015) 176-183.

[69] M.T. Carter, et al. ECS Trans. 50 (2012) 211-220.

[70] C. Hu, et al. Anal. Chem. 84 (2012) 3745-3750.

[71] S.W. Yun, J.R. Cha, M.S. Gong, Sens. Actuators B: Chem. 202 (2014) 1109-1116.

[72] A.S.G. Reddy, et al. Proc. Eng. 25 (2011) 120-123.

[73] J. Weremczuk, G. Tarapata, R.S. Jachowicz, Meas. Sci. Technol. 23 (2012) 014003.

[74] A. Rivadeneyra, et al. Sens. Actuators B: Chem. 195 (2014) 123-131.

[75] F. Molina-Lopez, Inkjet-Printed Multisensor Platform on Flexible Substrates for Environmental Monitoring, (Ph.D. thesis), EPFL, 2014 N° 6191.

[76] F. Molina-Lopez, D. Briand, N.F. de Rooij, Sens. Actuators B: Chem. 166-167 (2012) 212-222.

[77] F. Molina-Lopez, et al. in: Proceedings of the IEEE Sensors Conference, Taipei, Taiwan, (2012), pp. 220-223.

[78] F. Molina-Lopez, D. Briand, N.F. de Rooij, Org. Electron. 16 (2015) 139-147.

[79] S. Claramunt, et al. Sens. Actuators B: Chem. 187 (2013) 401-406.

[80] E. Danesh, et al. Anal. Chem. 86 (2014) 8951-8958.

[81] Rieu, et al. in: Proceedings of Eurosensors XXIX, Procedia Engineering, 2015.

[82] C. Peter, et al. Microsys. Technol. 18 (2012) 925-930.

[83] J. Courbat, et al. Sens. Actuators B: Chem. 143 (2009) 62-70.

[84] J. Courbat, et al. Proc. Eng. 25 (2011) 1329-1332.

[85] J. Courbat, et al. Sens. Actuators B: Chem. 160 (2011) 910-915.

[86] Z. Lin, et al. J. Electron. Packag. 135 (2013), article number 011001.

[87] J. Maklin, et al. Phys. Status Solidi B 245 (2008) 2335-2338.

[88] J. Kukkola, et al. J. Mater. Chem. 22 (2012) 17878-17886.

[89] W. Shen, Sens. Actuators B: Chem. 166-167 (2012) 110-116.

[90] C. Peter, J. Kneer, J. Wollenstein, Sens. Lett. 9 (2011) 1-5.