Flexible polyimide-based force sensor

Flexible polyimide-based force sensor

Sensors and Actuators A 173 (2012) 127–135 Contents lists available at SciVerse ScienceDirect Sensors and Actuators A: Physical journal homepage: ww...

2MB Sizes 9 Downloads 62 Views

Sensors and Actuators A 173 (2012) 127–135

Contents lists available at SciVerse ScienceDirect

Sensors and Actuators A: Physical journal homepage: www.elsevier.com/locate/sna

Flexible polyimide-based force sensor Jagoda A. Dobrzynska ∗ , Martin A.M. Gijs Laboratory of Microsystems, Ecole Polytechnique Fédérale de Lausanne, Station 17, CH-1015 Lausanne, Switzerland

a r t i c l e

i n f o

Article history: Received 14 July 2011 Received in revised form 29 October 2011 Accepted 7 November 2011 Available online 16 November 2011 Keywords: Force sensor Capacitive sensor Flexible sensor Polyimide Pattern transfer Capacitor model

a b s t r a c t We have realized a flexible force sensor, composed of four redundant capacitors, the operation of which is based on the measurement of a load-induced capacitance change. We use polyimide both as flexible substrate and as elastic dielectric between two levels of finger-shaped aluminum electrodes. In particular we have developed a technology for realization of two-level polyimide microstructures with gentle slopes to facilitate subsequent metallization processes. Thereby, we could improve step coverage and electrical contacting between the two metallization levels, as well as the mechanical stability of the sensor. The smooth polyimide slopes were obtained by combining lithographic resist-reflow techniques with dry etching procedures. We have analytically modeled the sensor’s capacitance and its force sensitivity. We have electrically characterized the capacitors using an impedance analyzer and obtained capacitances in the range of 130 pF and a typical force sensitivity of 0.5–1 fF/N, in excellent agreement with our model. © 2011 Elsevier B.V. All rights reserved.

1. Introduction The realization and application of microfabricated tactile sensors is well documented in the literature [1–3]. Technological advances have already provided small and thin sensors, with high spatial resolution of the sensing cells. A variety of applications deals with sensor arrays that can accurately measure hand forces [4–8]. Since measurement of contact forces is critical in robotics for determining the parameters of a gripper to grasp an object [9], micromachined force sensors are also nowadays investigated for robotic systems’ applications and artificial skin [3,10–13]. A force sensor can also be applied under the plantar surface of the foot for gait analysis, allowing evaluation of a patient’s health state [14] or measurement of the patient’s progress after orthopedic surgical intervention. Both stationary and ambulatory plantar force measuring devices are able to provide comprehensive information about the human movement. Nevertheless, wearable devices are cheaper and easier in employment, require less data processing, and allow long distance monitoring, during daily activities [15]. Following the first measurements of ground reaction forces (GRF) realized in the 19th century [16], numerous attempts have been made to develop portable or wearable gait analysis systems [17–24]. The most recent operate on strain gauge [22,25–28] and capacitive basis [18,19,21,23,29]; however, other solutions such as piezoelectric films [30–32], magneto-resistors [33,34], or air bladders [16,35] have also been reported. These force sensors are either built from

∗ Corresponding author. Tel.: +41 21 693 6596; fax: +41 21 693 5950. E-mail address: [email protected]fl.ch (J.A. Dobrzynska). 0924-4247/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.sna.2011.11.006

off-the-shelf components, or from centimeter- to millimeter-scaled elements that are machined in a standard workshop. To the best of our knowledge, clean-room batch microfabrication techniques have not been used before for plantar forces’ measurements for human gait analysis. However, silicon-based batch-fabricated tactile sensors have been reported. Kane et al. composed an array of piezoresistive stress sensing elements, constructed with a fully CMOS-compatible fabrication process, allowing integration of the sensing structures with digital control circuitry [36]. The sensors demonstrated linear responses to applied normal stress up to 35 kPa. Nevertheless, although these silicon-based force sensors offered miniaturization, they were not useful as tactile imagers, because of their hard and bulky plastic packaging and lack of force-overload protection. A capacitive sensor based on the electrostriction effect and proposed by Tseng et al. [37] exploited the dielectric particles that were encapsulated in a silicone rubber dielectric. The device was mounted on a rigid silicon substrate and, although the sensor was aimed for contact force measurements in orthodontia, no patient data were reported. Chu et al. developed a tactile sensor for application in robotic fingers and grippers, based on rigid silicon and glass substrates [38], allowing force measurements up to 10 mN. These authors suggested, but did not implement, the embedding of the silicon–glass sensor cells in a protective elastomer, as a convenient approach for the improvement of wear resistance and decrease of the sensors’ stress sensitivity. Furthermore, the connection of the sensor with the readout circuit was achieved with wire bonding, a delicate technique compromising the durability of the device. Hybrid piezoresistive tactile sensors composed of both rigid silicon posts and a flexible PI diaphragm, mimicking a


J.A. Dobrzynska, M.A.M. Gijs / Sensors and Actuators A 173 (2012) 127–135

Fig. 1. A conceptual view of the metallic electrode parts of the capacitive structure.

hair-like structure, have been reported by Hu et al. [39]. They revealed an improved durability and resistance to excessive load and were able to measure forces up to 3 mN. Nevertheless, flexibility of the sensor was limited due to a silicon-based packaging of the sensing cells. Batch-fabricated polymer-based flexible force sensors have already found applications in tactile imagers and artificial skins [10,12,13,40] or plantar pressure measuring systems for humanoid robots [11]. The usual compliant film substrates and materials used for these sensors are polydimethylsiloxane (PDMS) [13,40], parylene-C [41], and polyimide (PI) [11], the latter being commonly used as a dielectric interlayer in integrated circuits [42] or a flexible package material for sensors [43]. Polymer substrates can be fixed on the non-planar surfaces of robotic grippers, or can be used in other fields of application where flexibility and deformability of the substrate is necessary. In particular, recent artificial skin concepts have been based on fully flexible and deformable substrates [10–13,44] and advances in polymer technologies have given rise to the development of PI-based packaging of siliconbased sensors [5,11,43,45]. The compliant packaging provides to the sensor the desired flexibility, durability and overload protection. Hwang et al. proposed four thin metal strain gauges that were embedded in a ductile polymer substrate, creating a loadsensitive cell above which a flexible bump structure was placed

[11]. The sensor was used to measure the GRF of a humanoid robot and was able to measure loads up to 4 N, tolerating overloads without breakdown. Another flexible material for tactile applications is PDMS [10,12,13,46]. Lee et al. realized capacitive cells deposited on and embedded in PDMS layers, to build a tactile array for large area deployment [12,13]. The force range of this sensor was 10 mN, corresponding to a pressure of about 130 kPa. Polymer technology also offers the advantage of easy realization of three-dimensional (3D) microstructures or microstructures with inclined sidewalls, which can be of use for manufacture of optical lenses [47], microcavities [48], or suspended bridges and cantilevers [49]. So far, most reported techniques apply to the 3D structuring of photoresist (PR), and do less concern alternative resins such as PI. While piezoresistive [8,50], piezoelectric [51,52], inductive [6], and strain gauge principles [11] can be found in force sensing systems, the most common read-out scheme is capacitancebased [10,13,37,38,44,53–58]. A capacitive approach is preferred due to numerous advantages when compared to resistive and piezoresistive systems [59]: capacitive sensors have high sensitivity, better temperature performance, are less sensitive to drift, and have low power consumption, a feature especially essential for autonomous portable applications. Furthermore, most of the capacitive sensors measure the displacement of a membrane, not its stress, and therefore are intrinsically more robust. In this paper, we present PI-based capacitive flexible force sensors that are realized using a clean room batch-type process. Our sensor has the following desirable features: (i) an excellent durability and resistance to forces in the kN range, (ii) flexibility of the sensor due to elastic packaging, and (iii) reliability by choice of a robust bonding scheme. In our technology, we propose two methods for the fabrication of oblique profiles in PI, which facilitate subsequent thin film metallization processes, electrical contacting between different metallization levels, as well as the mechanical stability of the sensor. The smooth PI slopes are obtained by combining lithographic resist-reflow techniques [47,49] with dry

Fig. 2. Process flow for the fabrication of the capacitive sensor: (a) sputter sacrificial W/Al0, (b) spin coat PI1, (c) sputter and pattern Ti/Al1 bottom electrode, (d) spin coat PI2 dielectric layer, (e) realization of a smooth PR slope, (f) dry etch PI2, (g) sputter and pattern Ti/Al2 top electrode, (h) spin coat PI3, (i) dry etch PI3 to open contact pads and (j) anodic dissolution of sacrificial Al0 layer.

J.A. Dobrzynska, M.A.M. Gijs / Sensors and Actuators A 173 (2012) 127–135


Fig. 3. X-ray photoelectron spectroscopy (XPS) analysis of the Al contact pad surface, as prepared (blue line), and treated with an ion beam (red line) to remove the top oxidized layer. Two distinct natures of oxidation were examined: (a) native oxide: grown under atmospheric conditions and (b) oxide: formed in an oxygen plasma process. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

Fig. 4. (a) Flexible sensor released from the silicon support wafer and bonded to a PCB; (b) magnified view of the connection pads area. (c) SEM picture of the gold bump bonding: 3 gold bumps on the surface of the square Al contact pad, form the through-hole contact with the PCB. The scale bar is 100 ␮m. (d) Cross-section view of the through-hole bonded pads.

Fig. 5. SEM cross-section micrographs (1) after realization of the smooth slope in PI2 and (2) after Al2 metal coverage, for two different photoresists used as a mask: (a) AZ ECI 3027 and (b) AZ 9260. The scale bars correspond to 5 ␮m. (c) SEM cross-section micrograph of the finalized device. The scale bar corresponds to 10 ␮m.


J.A. Dobrzynska, M.A.M. Gijs / Sensors and Actuators A 173 (2012) 127–135

Fig. 6. Frequency characteristics of the four redundant capacitors of one sensor: (a) impedance and phase angle, and (b) capacitance derived from the imaginary part of the complex impedance, negative values therefore indicating inductive behavior. The overlapping curves indicate the high reproducibility of the fabrication process.

etching procedures. Finally, we present an analytical model and electrical characterization of our sensor. 2. Materials and methods 2.1. Sensor design and fabrication Fig. 1 illustrates a schematic diagram of the metal electrode parts of our sensor: two levels of finger-like microstructures form four redundant capacitors (C1–C4) with common top electrode. The finger-like geometry of the sensor is chosen to maintain flexibility of the sensor, when it is deformed during force application on nonflat surfaces, thereby preventing deformation and cracking of the relatively large electrode surface, which could be a problem when using a parallel full-plate capacitor. Contact pads are designed on the lower metal layer, which is electrically connected to the top layer via a tapered pad. Consequently, mechanical stability of the sensor is maintained and bonding on the contact pads is possible, even when the top electrode is displaced with regard to the bottom electrode. The electrode structure is embedded in PI, except for openings kept on the contact pads. Each capacitor has an area of 5 mm × 5 mm, and the thickness of the insulation layer between

top and bottom electrode in the absence of a load is fixed at 5 ␮m. The cell deforms under a normally applied force, decreasing the thickness of the insulation layer and increasing the capacitance value. The process flow of the sensor is shown in Fig. 2. We start by sputtering on a silicon substrate a 500 nm conductive layer of tungsten, followed by deposition of a sacrificial 1 ␮m Al0 layer (Fig. 2a) needed for anodic dissolution of the finalized flexible structure from the silicon support. The 5 ␮m thick layer PI1 is spin-coated and baked (Fig. 2b), and a low stress PI (PI2611, HD Microsystems) is chosen in order to avoid buckling of the sensor structure after release from the silicon support wafer. The lower 500 nm thick aluminum electrode layer (Al1) is sputtered following an oxygen plasma surface activation step of the PI1 layer after deposition of a 50 nm film of Ti for adhesion. The electrodes are patterned in a dry etching process with Cl2 /BCl3 chemistry (Fig. 2c). The 5 ␮m dielectric layer PI2 is spin-coated, baked and patterned in such a way, that smooth slopes in PI2 are obtained (Fig. 2d). A photolithographic process is used to structure the PR in a tapered shape (Fig. 2e), followed by an anisotropic plasma etch. Two distinct PRs are tested: (i) AZ ECI 3027 (AZ Electronic Materials), a PR which after exposure and development exhibits oblique sidewall topography, and

Fig. 7. Two-dimensional simulation results of the electrical field and potential for a finger-like electrode capacitive sensor. The geometrical parameters are defined in Sakurai’s formula.

J.A. Dobrzynska, M.A.M. Gijs / Sensors and Actuators A 173 (2012) 127–135


Fig. 8. Numerically simulated capacitance vs. dielectric thickness, and its fit with Eq. (1).

(ii) AZ 9260 (AZ Electronic Materials), a PR with vertical sidewalls, which we melt after the development, on a hotplate at 130 ◦ C. Next, the 5 ␮m thick tapered PR pattern is transferred into PI2 by an anisotropic oxygen plasma etch (Fig. 2f). Because of the very poor selectivity of the PR mask with respect to the etched PI2 layer, the profile of the PR is directly transmitted into the PI2 substrate. The tapered slope allows conformal metallization by a 50 nm Ti adhesion layer, followed by a 500 nm Al2 top electrode layer, which is patterned in a dry etching process (Fig. 2g). The process continues by depositing the 5 ␮m thick PI3 layer (Fig. 2h), and deposition and structuring of an amorphous Si mask for defining the contact pad openings during a subsequent oxygen plasma dry etching step (Fig. 2i). Finally, the structures are released from the rigid Si wafer by anodic dissolution of the sacrificial Al0 [60] (Fig. 2j). After microfabrication, the sensors are annealed at a temperature of 100 ◦ C for 30 min, at a 3 kN load, in order to reduce the influence of eventual material meta-stabilities, cracks or irregularities, both in the PI and the electrodes. 2.2. Sensor packaging For the connection of the sensor to the measurement system, several bonding techniques are investigated, such as conductive epoxy-based adhesives, soldering pastes, or polymeric anisotropically conductive adhesives. However, due to the extensive oxygen plasma dry etching processes, the oxidation of the aluminum contact pads imposes a serious connection problem. The results of the X-ray photoelectron spectroscopy (XPS) analysis of the aluminum pad surface are shown in Fig. 3. In the analysis, two samples of aluminum pads are examined. The first sample (Fig. 3a) is oxidized in atmospheric conditions resulting in a native aluminum oxide layer. The second one (Fig. 3b) is treated with an oxygen plasma etch, which is long enough to open the metallic contact pads covered by the top polymeric layer. In order to remove the excess of the surface aluminum oxide layer, both samples are treated in situ with an ion beam (Ar+, 3 kV, 2 min). Based on the calculations of Carlson [61] and Strohmeier [62], the analysis reveals an oxide layer that is twice as thick (6.4 nm), compared with the native oxide (3.2 nm) on top of the connection pad. As the oxide layer is difficult to be removed with standard microfabrication processes, we prefer to fissure this layer and make the connection with the underlying metallic aluminum using mechanical action via a gold bump bonding technique [45], attaching the sensor reliably to a printed circuit board (PCB), and maintaining the flexibility of the package. Fig. 4a–d shows a sensor, obtained after release from the silicon support, and attached to the PCB, connecting sensory pads with a golden stud.

Fig. 9. Conceptual view of the experimental setup for testing the force sensor: (a) gripper, (b) screw, (c) bearing, (d) blocks, (e) flexible force sensor, (f) PCB, (g) connection to impedance analyzer, (h) load cell and (i) connection to a PC.

Fig. 5a and b are Scanning Electron Microscopy (SEM) graphs showing the smooth PI2 slope after the PR structuring/dry etching step and after sputtering layer Al2, for the two methods proposed. Fig. 5c is a cross-section SEM graph of the finalized sensor stack embedded in three polyimide layers before release from the rigid support wafer, illustrating the connection between the top and bottom electrode via a tapered Al slope. The cross-section samples are prepared by embedding the sensor in a hard epoxy, and polishing with a tripod on a diamond paper with a grain size of up to 500 nm (Allied High Tech Products Inc.). 3. Modeling, electrical characterization and discussion Fig. 6 shows the electrical characterization of the four unloaded capacitors of a single sensor, as measured, at stable laboratory conditions with room temperature and 70% relative humidity, by an Agilent 4294A impedance analyzer. The initial measured

Fig. 10. Sensor response to applied experimental loads: capacitance vs. frequency characteristics.


J.A. Dobrzynska, M.A.M. Gijs / Sensors and Actuators A 173 (2012) 127–135

Fig. 11. Measurement results: output capacitance at 5 kHz vs. applied compressive load for four capacitors of the same sensor during loading (squares) and unloading cycle (circles).

capacitances were in the range of 120–140 pF, and the four curves are superimposed. The measured cut-off frequency of the device is 10 MHz. Finite Element Method (FEM) analysis (COMSOL Multiphysics 3.5a) is used to compare the modeled capacitance to its experimentally measured values. The simulated sensor’s geometry is illustrated in Fig. 1 and the structure’s dimensions are those mentioned in Section 2.1. For the analysis, the dielectric material constant of the polymer is chosen εr = 4 [63]. The simulation results are shown in Fig. 7 (with indication of t, s, w, h geometrical parameters) and represent two-dimensional (2-D) analysis along the cross-section plane, similar to the one presented in Fig. 5c. The calculated theoretical value of the as-made capacitor is 125 pF, which

is in good agreement with the experiment. We also used the FEM analysis for the simulation of the effect of polymer thickness h. The results are shown in Fig. 8. The capacitance change as a function of the dielectric’s thickness is in excellent agreement with a modified Sakurai’s formula [64] (1).

C∼ = nε0 εr

+a4 0.83

a1 1.15

t h

w h

 t 0.222

+ a2 2.8


+ 2 a3 0.03

 t 0.222   h 1.34 

+ a5 0.07




w h (1)

where C is the capacitance between the top and bottom electrode, n is the number of electrode fingers, ε0 is the vacuum permittivity, h

Fig. 12. Loading experiments of four capacitors (C1–C4) in the lower force range of 0–200 N.

J.A. Dobrzynska, M.A.M. Gijs / Sensors and Actuators A 173 (2012) 127–135


Table 1 Comparison of capacitive force detecting sensors. Device


Maximum force limit

Cut-off/resonance frequency

Sensory surface flexibility

This work

0.5–1 fF/N 2–4 fF/1 MPa ∼0.8%/kN

3 kN 120 MPa

10–20 MHz


Cheng [10]

290 fF–1.78 pF/100 kPaa 2.27–0.03%/kPa

5N 430 kPa

13 kHz


Lee [12,13]

5–6 fF/mNa 43–48 fF/100 kPaa 2.9%/mN

10–40 mN 130–250 kPa

Not reported


Cotton [40]

13–17 fF/100 kPa

∼5 Na 160 kPa

>32 kHz


Kothari [19]

17 pF/100 kPaa

3 MPa

100–200 kHz


Tseng [37]

26.9 pF/100 kPa

50 mN

Not reported


Chu [38]

13 fF/mN

10 mN

162 Hz



Values interpreted from figures or estimated from the available in the publication data.

is the dielectric thickness, w is the width of a single electrode finger, s is the spacing between single electrode fingers, t is the thickness of a single electrode, L is the length of electrode fingers, and a1 –a5 are fixed fitting parameters of order one, which take into account that the two capacitor electrodes have finger-like structures (the original Sakurai formula is only modeling the stray capacitance between a finger-like electrode and the continuous plane counter electrode). Force sensing experiments are done with the measurement setup shown in Fig. 9. A self-centering, self-gripping puller with a ball-bearing ended screw is used to compress the force sensor locked in between stiff blocks and an iLoad Pro Digital load cell (LoadStar Sensors), the latter being used to measure the applied force. The sensor’s response was found to be susceptible to the nature of the material of the blocks, especially if these have conductive metallic surfaces. When a conductive object is in proximity with the surface of the sensor, it disturbs the capacitive coupling between the electrodes. Indeed, when the material is electrically conductive and floating, it acts as an additional capacitive plate and the measured capacitance C increases; e.g. when a conductive steel plate is placed at 150 ␮m distance to the sensor, its initial capacitance C0 = 130.9 pF increases by ∼1.5 pF. On the contrary, when the sensor is close to a grounded metal plate, the measured capacitance C decreases. For our loading experiments, we avoid metals in proximity of the capacitive sensor and use polymethylmethacrylate (PMMA) as material for the blocks. Fig. 10 shows the sensor’s capacitive response to two applied loads. The measurements of the relative capacitance change as a function of frequency show that the capacitors respond similarly at all frequencies outside the range of the resonance. A frequency of 5 kHz was chosen for further electrical characterization of the sensor. Fig. 11 shows the results of the measured capacitance vs. applied force in the 0–1 kN range. In order to minimize errors in the measurement, the average value from 5 consecutive sample measurements is taken for every data point and the error bars indicate the standard deviation from the mean value. Both sensor loading and unloading characteristics are obtained, revealing no hysteresis effect. Moreover, the sensor response is nearly linear over the whole measurement range. The value of the experimentally measured capacitance is in excellent agreement with the computer simulations from Fig. 8. The deviation from linearity in the lower force range (up to 100 N) could be due to the stress–strain behavior of the compressed polymer, or the presence of microcracks in the aluminum film, an effect that has been already observed for mechanically stressed platinum films [65]. A minor difference of initial capacitance values between the four capacitors can be

noticed, most probably due to the geometrical design of the electrical leads. From dedicated loading experiments in the lower force range, illustrated in Fig. 12, we find a lower limit of detection of 100 N. The force sensitivity of the sensor C/F = (C/h) · (h/F), is calculated from (1) and (2): F = F0

h h0


b2 C −b1 + = F [F0 + F] F b3 + F

 0.222 F0

  F0 + F 0.34 F0

F0 + F

b4 −1 + F


  F0 + F 1.118 F0



where h0 and h is the dielectric’s initial thickness and its change, and b1 –b4 and F0 are geometrical and material constants defined as: b1 = 1.15nε0 εr wLa1 h0


b2 = 2.8nε0 εr t 0.222 La2 h0


b3 = nε0 εr s−1.34 L(0.06wa3 + 1.66ta4 )h0 b4 = 0.14nε0 εr t 0.222 s−1.34 La5 h0




F0 = EA and E is the dielectric’s Young’s modulus.

Fig. 13. Sensitivity of the sensor C/F vs. applied force: as measured (points) and calculated from Eq. (3) (curves), for various values of polymer’s Young’s modulus E.


J.A. Dobrzynska, M.A.M. Gijs / Sensors and Actuators A 173 (2012) 127–135

Fig. 14. Measured change in capacitance vs. force applied to the sensor when compressed (a) on a flat surface and (b) in-between convex and concave cylindrical surfaces with radius R.

Fig. 13 illustrates the sensitivity of the sensor over the whole measured force range. For forces above 200 N, we find a sensitivity of C/F ∼ 0.5–1 fF/N, which corresponds to ∼2–4 fF/MPa, in agreement with similarly designed sensors [40]. For a given elastic modulus E, the calculated sensitivity (3) is compared with the measurement results in Fig. 13. The theoretical prediction using a PI’s modulus value of E = 7 GPa [63] shows a good match with the experimental data. Table 1 compares our sensor with capacitive force-detecting sensors reported in literature. The maximum applicable force on our sensor exceeds that of similar flexible sensors by a factor of minimum 100, resulting in a more suitable range for large force applications. Fig. 14 compares the response of the flat sensor to the situation when it is bent, by placing it in-between convex and concave cylindrical surfaces with various radii R. A minor decrease in capacitance is obtained, which is due to differences between the normal force applied to the surface of the capacitor in the flat and bent case. 4. Conclusions and outlook A fully flexible, capacitive force sensor for large force measurements has been developed, by introducing a tapered slope clean room fabrication process in PI. Robust packaging of the sensor is obtained by embedding all metal electrode parts in PI, and a reliable connection with a PCB is realized by using a gold bump bonding technique. The experimental capacitive values are in agreement with theoretical predictions. The sensor has high strength and durability, when operated in the 1 kN force range, exhibits linearity and absence of hysteresis, a detection limit of 100 N, as well as an acceptable sensitivity of 0.5–1 fF/N. The sensor can continually withstand higher loads, in excess of 3 kN, without damage. By changing the physical parameters of the sensor (i.e. dielectric thickness, polymer’s elasticity, plate size) the sensor’s characteristics can be tailored to the application. For robotic applications, with metallic grippers or metallic objects to grip, the susceptibility of the sensor to conductive surfaces resulting in parasitic stray capacitance can be handled by surrounding the sensor with appropriate guard or shield electrodes. In the future, we intend to modify the elasticity of the inter-electrode polymer in order to exploit our sensor for shear force detection, and interface it with appropriate electronic circuitry for the automatized measurement of plantar forces for human gait analysis applications. Acknowledgements This work has been carried out thanks to the equipment of the EPFL-CMI and the funding from the Swiss National Science Foundation (project 32003B-12042). The authors would like to thank C. Hibert and G. Clerc for help with the fabrication process, C. Vallotton for sample preparation methods, V. Laporte for the XPS analysis,

and K. Aminian, X. Crevoiser, and H. Rouhani for useful suggestions and discussions regarding the sensor’s design and application.

References [1] J.J. Boland, Flexible electronics: within touch of artificial skin, Nature Materials 9 (10) (2010) 790–792. [2] P. Dario, Tactile sensing: technology and applications, Sensors and Actuators A: Physical 26 (1–3) (1991) 251–256. [3] P.P.L Regtien, Tactile imaging, Sensors and Actuators A: Physical 31 (1–3) (1992) 83–89. [4] D.J. Beebe, et al., A silicon-based tactile sensor for finger-mounted applications, IEEE Transactions on Biomedical Engineering 45 (2) (1998) 151–159. [5] J. Engel, J. Chen, L. Chang, Development of polyimide flexible tactile sensor skin, Journal of Micromechanics and Microengineering 13 (3) (2003) 359–366. [6] N. Futai, K. Matsumoto, I. Shimoyama, A flexible micromachined planar spiral inductor for use as an artificial tactile mechanoreceptor, Sensors and Actuators A: Physical A111 (2-3) (2004) 293–303. [7] T.R. Jensen, R.G. Radwin, J.G. Webster, A conductive polymer sensor for measuring external finger forces, Journal of Biomechanics 24 (9) (1991) 851–858. [8] D.J. Beebe, et al., A silicon force sensor for robotics and medicine, Sensors and Actuators A: Physical A50 (1–2) (1995) 55–65. [9] R.S. Johansson, G. Westling, Roles of glabrous skin receptors and sensorimotor memory in automatic control of precision grip when lifting rougher or more slippery objects, Experimental Brain Research 56 (3) (1984) 550–564. [10] M.Y. Cheng, et al., A flexible capacitive tactile sensing array with floating electrodes, Journal of Micromechanics and Microengineering 19 (11) (2009) p115001. [11] E.-S. Hwang, J.-H. Seo, Y.-J. Kim, A polymer-based flexible tactile sensor for both normal and shear load detections and its application for robotics, Journal of Microelectromechanical Systems 16 (3) (2007) 556–563. [12] H.-K. Lee, S.-I. Chang, E. Yoon, A flexible polymer tactile sensor: fabrication and modular expandability for large area deployment, Journal of Microelectromechanical Systems 15 (6) (2006) 1681–1686. [13] H.-K. Lee, et al., Normal and shear force measurement using a flexible polymer tactile sensor with embedded multiple capacitors, Journal of Microelectromechanical Systems 17 (4) (2008) 934–942. [14] H. Rouhani, et al., Ambulatory assessment of 3D ground reaction force using plantar pressure distribution, Gait and Posture (2010). [15] J. Favre, et al., Evaluation of a mixed approach combining stationary and wearable systems to monitor gait over long distance, Journal of Biomechanics 43 (11) (2010) 2196–2202. [16] M. Marey, De la locomotion terrestre chez les bipedes et les quadrupedes, Journal de l’Anatomie et de la Physiologie 9 (1873) 42–80. [17] P. Hargreaves, J.T. Scales, Clinical assessment of gait using load measuring footwear, Acta Orthopaedica Scandinavica 46 (6) (1975) 877–895. [18] T.S. Holden, R.W. Muncey, Pressures on the human foot during walking, Australian Journal of Applied Science 4 (405) (1953). [19] M. Kothari, et al., Capacitive sensors for measuring the pressure between the foot and shoe, in: Annual International Conference of the IEEE Engineering in Medicine and Biology Society, 1988. [20] R.C. Levin, J.I. Molinder, S.S. Serdengecti, A system for the measurement of vertical foot forces, in: 2nd Conference on the Systems and Devices for the Disabled, 1976. [21] S. Miyazaki, A. Ishida, Capacitive transducer for continuous measurement of vertical foot force, Medical and Biological Engineering and Computing 22 (4) (1984) 309–316. [22] S. Miyazaki, H. Iwakura, Foot-force measuring device for clinical assessment of pathological gait, Medical and Biological Engineering and Computing 16 (4) (1978) 429–436. [23] A. Santic, V. Bilas, I. Lackovic, A system for measuring forces in the legs and crutches from ambulatory patients, in: International Conference of the IEEE Engineering in Medicine and Biology Society, IEEE, Piscataway, NJ, USA, 1997. [24] G. Spolek, et al., Ambulatory-force measurement using an instrumented-shoe system, Experimental Mechanics 15 (7) (1975) 271–274.

J.A. Dobrzynska, M.A.M. Gijs / Sensors and Actuators A 173 (2012) 127–135 [25] B. Davis, et al., A device for simultaneous measurement of pressure and shear force distribution on the plantar surface of the foot, Journal of Applied Biomechanics 14 (1) (1998) 93–104. [26] A. Faivre, et al., Instrumented shoes for pathological gait assessment, Mechanics Research Communications 31 (5) (2003) 627–632. [27] T. Liu, Y. Inoue, K. Shibata, Wearable force sensor with parallel structure for measurement of ground-reaction force, Measurement 40 (6) (2007) 644–653. [28] H.S. Ranu, Miniature load cells for the measurement of foot-ground reaction forces and centre of foot pressure during gait, Journal of Biomedical Engineering 8 (2) (1986) 175–177. [29] T. Salpavaara, J. Verho, J. Lekkala, Capacitive insole sensor for hip surgery rehabilitation, in: Second International Conference on Pervasive Computing Technologies for Healthcare (PervasiveHealth), 2008. [30] F. Akhlaghi, M. Pepper, In-shoe biaxial shear force measurement: the Kent shear system, Medical and Biological Engineering and Computing 34 (4) (1996) 315–317. [31] T.S. Gross, R.P. Bunch, Measurement of discrete vertical in-shoe stress with piezoelectric transducers, Journal of Biomedical Engineering 10 (3) (1988) 261–265. [32] M.A. Razian, M.G. Pepper, Design, development, and characteristics of an inshoe triaxial pressure measurement transducer utilizing a single element of piezoelectric copolymer film, IEEE Transactions on Neural Systems and Rehabilitation Engineering 11 (3) (2003) 288–293. [33] J.W. Tappin, et al., Method of measuring ‘shearing’ forces on the sole of the foot, Clinical Physics and Physiological Measurement 1 (1) (1980), 83. [34] R.B. Williams, Triaxial force transducer for investigating stresses at the stump/socket interface, Medical and Biological Engineering and Computing 30 (1) (1992) 89–96. [35] K. Kong, M. Tomizuka, A gait monitoring system based on air pressure sensors embedded in a shoe, IEEE/ASME Transactions on Mechatronics 14 (3) (2009) 358–370. [36] B.J. Kane, M.R. Cutkosky, G.T.A. Kovacs, A traction stress sensor array for use in high-resolution robotic tactile imaging, Journal of Microelectromechanical Systems 9 (4) (2000) 425–434. [37] F.-G. Tseng, C.-S. Yang, L.-C. Pan., An elastomeric tactile sensor employing dielectric constant variation and applicable to orthodontia, in: 17th IEEE International Conference on Micro Electro Mechanical Systems, IEEE, Piscataway, NJ, USA, 2004. [38] Z. Chu, P.M. Sarro, S. Middelhoek, Silicon three-axial tactile sensor, Sensors and Actuators A: Physical 54 (1–3) (1996) 505–510. [39] Y. Hu, et al., Bioinspired 3-D Tactile Sensor for Minimally Invasive Surgery, Journal of Microelectromechanical Systems 99 (2010) 1–9. [40] D. Cotton, I.M. Graz, S.P. Lacour, A multifunctional capacitive sensor for stretchable electronic skins, IEEE Sensors Journal 9 (12) (2009) 2008–2009. [41] Y. Hasegawa, et al., Fabrication of a wearable fabric tactile sensor produced by artificial hollow fiber, Journal of Micromechanics and Microengineering 18 (8) (2008) p085014. [42] D. McHerron, E. Perfecto, T. Redmond, MCM-D polyimide applications at IBM, in: National Electronic Packaging and Production Conference, NEPCON East ‘94, Reed Exhibition Companies, Stamford, CT, USA, 1994. [43] D.J. Beebe, D.D. Denton, Flexible polyimide-based package for silicon sensors, Sensors and Actuators A: Physical 44 (1) (1994) 57–64. [44] G. Kita, et al., Development of fabric force sensor for detecting normal and lateral force by applying umbonal fiber, in: Transducers 2009, IEEE Computer Society, Denver, CO, USA, 2009. [45] Y. Moser, Miniaturized flexible temperature sensor, Journal of Microelectromechanical Systems 16 (6) (2007) 1349–1354. [46] A. Shashank, et al., Design, simulation and fabrication of a low cost capacitive tactile shear sensor for a robotic hand, in: Conference of the IEEE Engineering in Medicine and Biology Society, IEEE Computer Society, Minneapolis, MN, USA, 2009. [47] H. Yang, et al., Micro-ball lens array modeling and fabrication using thermal reflow in two polymer layers, Journal of Micromechanics and Microengineering 14 (2) (2004) 277–282. [48] Y.-J. Chang, K. Mohseni, V.M. Bright, Fabrication of tapered SU-8 structure and effect of sidewall angle for a variable focus microlens using EWOD, Sensors and Actuators A: Physical 136 (2) (2007) 546–553. [49] S.C. Saha, et al., Tuning of resist slope with hard-baking parameters and release methods of extra hard photoresist for RF MEMS switches, Sensors and Actuators A: Physical 143 (2) (2008) 452–461. [50] A. Tibrewala, Development, fabrication and characterization of a 3D tactile sensor, Journal of Micromechanics and Microengineering 19 (12) (2009), 125005-125005.


[51] S. Kärki, et al., Development of a piezoelectric polymer film sensor for plantar normal and shear stress measurements, Sensors and Actuators A: Physical 154 (1) (2009) 57–64. [52] T. Polster, M. Hoffmann, Aluminum nitride based 3D, piezoelectric, tactile sensor, Procedia Chemistry 1 (1) (2009) 144–147. [53] R.A. Boie, Capacitive impedance readout tactile image sensor, in: Proceedings of the IEEE International Conference on Robotics and Automation, 1984. [54] T.A. Chase, C.L. Ren., A thin-film flexible capacitive tactile normal/shear force array sensor, in: IECON ‘95 – 21st Annual Conference on IEEE Industrial Electronics, 1995. [55] K.J. Chun, K.D. Wise., A high-performance silicon tactile imager based on a capacitive cell, in: First World Conference on Robotics Research, Robotics Int. SME, Dearborn, MI, USA, 1984. [56] H.-K. Kim, et al., Transparent and flexible tactile sensor for multi touch screen application with force sensing, in: Transducers 2009, IEEE, Piscataway, NJ, USA, 2009. [57] P. Peng, et al., Novel MEMS stiffness sensor for force and elasticity measurements, Sensors and Actuators A: Physical 158 (1) (2010) 10–17. [58] F. Zhu, J.W. Spronck, A capacitive tactile sensor for shear and normal force measurements, Sensors and Actuators A: Physical 31 (1–3) (1992) 115–120. [59] R. Puers, Capacitive sensors: when and how to use them, Sensors and Actuators A: Physical (37–38) (1993) 93–105. [60] S. Metz, Partial release and detachment of microfabricated metal and polymer structures by anodic metal dissolution, Journal of Microelectromechanical Systems 14 (2) (2005) 383–391. [61] T.A. Carlson, G.E. McGuire, Study of the X-ray photoelectron spectrum of tungsten-tungsten oxide as a function of thickness of the surface oxide layer, Journal of Electron Spectroscopy and Related Phenomena 1 (2) (1972) 161–168. [62] B.R. Strohmeier, An ESCA method for determining the oxide thickness on aluminum alloys, Surface and Interface Analysis 15 (1990) 51–56. [63] HD Microsystems, Liquid Polyimides and PBO Precursors, 14.07.2011. Available from: http://hdmicrosystems.com. [64] T. Sakurai, K. Tamaru, Simple formulas for two- and three-dimensional capacitances, IEEE Transactions on Electron Devices ED-30 (2) (1983) 183–185. [65] E. Belloy, A. Sayah, M.A.M. Gijs, Micromachining of glass inertial sensors, Journal of Microelectromechanical Systems 11 (1) (2002) 85–90.


Jagoda A. Dobrzynska received her M.Sc. Eng. degree in Automation and Industrial Robotics in 2006 from Warsaw University of Technology, Poland. From 2006 to 2007 she was with Sener Ingeniería y Sistemas, Spain, working on industrial engineering design of railways, airports, shipyards, and power plants. She was a European Commission’s Marie Curie fellow in Numerical Methods Laboratory, Romania, from 2007 to 2009. She is currently working toward the Ph.D. degree in Microsystems at the Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland. Her primary research interests are in the development of new clean room microfabrication techniques and process engineering optimization. Martin A.M. Gijs received his degree in physics in 1981 from the Catholic University of Leuven, Belgium, and his Ph.D. degree in physics at the same university in 1986. He joined the Philips Research Laboratories in Eindhoven, The Netherlands, in 1987. He joined EPFL in 1997 as an associate professor, and was promoted to full professor in 2000. Professor Gijs has authored 200 peer-reviewed articles in international journals and holds about twenty patents. He has a strong research experience in the development of extremely thin polyimide-foil based inductive sensors, transformers and temperature sensors. His interests are in developing new microfabrication technologies for microsystems fabrication in general, and the development and use of microsystems technologies for microfluidic and biomedical applications in particular.